Highlights of Spanish Astrophysics IV

Highlights of Spanish Astrophysics IV Proceedings of the seventh Scientific Meeting of the Spanish Astronomical Society (SEA), held in Barcelona, Spain, September 12-15, 2006 Edited by Francesca Figueras Departament d’Astronomia i Meteorologia, Universitat de Barcelona-IEEC Margarita Hernanz Institut de Ciències de l’Espai, CSIC-IEEC Josep Miquel Girart Institut de Ciències de l’Espai, CSIC-IEEC Carme Jordi Departament d’Astronomia i Meteorologia, Universitat de Barcelona-IEEC Contents Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . IX Organizing Committees and Sponsors . . . . . . . . . . . . . . . . . . . . . . . . . . XI List of participants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . XIII Session I Spain in ESO Youth, accretion, and mass loss at the end of the main sequence F. Comerón . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 The European Extremely large Telescope P. Dierickx . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 Gamma-ray bursts: lighthouses of the Universe J. Gorosabel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 The VIMOS VLT Deep Survey (VVDS) R. Pelló, O. Le Fèvre, C. Adami, M. Arnaboldi, S. Arnouts, S. Bardelli, M. Bolzonella, A. Bongiorno, M. Bondi, D. Bottini, G. Busarello, A. Cappi, S. Charlot, P. Ciliegi, T. Contini, S. Foucaud, P. Franzetti, B. Garilli, I. Gavignaud, L. Guzzo, O. Ilbert, A. Iovino, F. Lamareille, V. Le Brun, D. Maccagni, B. Marano, C. Marinoni, G. Mathez, A. Mazure, H.J. McCracken, Y. Mellier, B. Meneux, P. Merluzzi, R. Merighi, S. Paltani, J.P. Picat, A. Pollo, L. Pozzetti, M. Radovich, V. Ripepi, D. Rizzo, R. Scaramella, M. Scodeggio, L. Tresse, G. Vettolani, A. Zanichelli, G. Zamorani, E. Zucca . . . . . . . . . . . . . . . . . . 41 Session II Science with GTC VI Contents Galaxy Surveys in the Era of Large Ground-Based Observatories R. Guzmán . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51 The GTC 10m telescope: Getting ready for First Light J.M. Rodrı́guez Espinosa, GTC Project . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63 OSIRIS: Status and Science J. Cepa, M. Aguiar, E.J. Alfaro, J. Bland-Hawthorn, H.O. Castañeda, F. Cobos, S. Correa, C. Espejo, A. Farah, A.B. Fragoso-López, J.V. Gigante, F. Garfias, J.J. González, V. González-Escalera, J.I. González-Serrano, B. Hernández, A. Herrera, C. Militello, L. Peraza, R. Pérez, J.L. Rasilla, B. Sánchez, M. Sánchez-Portal, C. Tejada . . . . . . 71 EMIR, the GTC NIR multi-object imager-spectrograph F. Garzón, D. Abreu, S. Barrera, S. Becerril, L.M. Cairós, J.J. Dı́az, A.B. Fragoso-López, F. Gago, R. Grange, C. González, P. López, J. Patrón, J. Pérez, J.L. Rasilla, P. Redondo, R. Restrepo, P. Saavedra, V. Sánchez, F. Tenegi, M. Vallbé . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81 CanariCam: Instrument Status and Frontier Science C.M. Telesco . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91 Session III S.E.A. prizes Radiative Transfer in Molecular Lines. Astrophysical Applications A. Asensio Ramos . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105 The star formation history of early-type galaxies as a function of environment P. Sánchez-Blázquez . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117 Session IV Galaxies and cosmology Galaxy Evolution in Galaxy Clusters: Diffuse Light in the Virgo Cluster J. Alfonso L. Aguerri . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 131 The quest for obscured AGN at cosmological distances: Infrared Power-Law Galaxies A. Alonso-Herrero, J.L. Donley, G.H. Rieke, J.R. Rigby, P.G. Pérez-González . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143 Contents VII AMIGA: A New Model of Galaxy Formation and Evolution A. Manrique on behalf of the AMIGA collaboration . . . . . . . . . . . . . . . . . . . 157 The innermost regions of Active Galactic Nuclei – from radio to X-rays E. Ros, M. Kadler, , S. Kaufmann, Y.Y. Kovalev, J. Tueller, K.A. Weaver . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165 Gaussian analysis of the CMB with the smooth tests of goodness of fit R.B. Barreiro, J.A. Rubiño-Martı́n, E. Martı́nez-González . . . . . . . . . . . . 177 Dark matter in galaxy clusters N. Benı́tez . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185 Cosmology with the Largest Scale Structures: Probing Dark Energy F. J. Castander, the Dark Energy Survey Collaboration . . . . . . . . . . . . . . . 193 Observational cosmology at high redshift A. Fernández-Soto . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201 An Hα approach to the evolution of the galaxy population of the universe J. Gallego, V. Villar, S. Pascual, J. Zamorano, K. Noeske, D.C. Koo, P.G. Pérez-González, G. Barro . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209 Session V The Galaxy and its components Multi-wavelength Astronomy and the unidentified γ-ray sources J. Martı́-Ribas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 219 The disc and plane of the Milky Way in the Near Infared A. Cabrera-Lavers, M. López-Corredoira, F. Garzón, P.L. Hammersley, C. González-Fernández, B. Vicente . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231 AGB Stars: Nucleosynthesis and Open Problems I. Domı́nguez, C. Abia, S. Cristallo, P. de Laverny, O. Straniero . . . . . . . 239 Studying galaxy formation and evolution from Local Group galaxies. C. Gallart . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 247 Gaia: A major step in the knowledge of our Galaxy J. Torra on behalf of the Gaia Group . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 255 VIII Contents Cepheus A, a laboratory for testing and opening new theories on high-mass star formation J.M. Torrelles, N.A. Patel, S. Curiel, G. Anglada, J.F. Gómez . . . . . . . . 263 Session VI Sun and planetary systems A Look into the Guts of Sunspots L.R. Bellot Rubio . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 271 Earth-like Exoplanets. Darwin: Stellar Targets and Precursor Science C. Eiroa, M. Fridlund, L. Kaltennegger, A. Stankov . . . . . . . . . . . . . . . . . 279 How the comet 9P/Tempel 1 has behaved before, during and after the Deep Impact event L.M. Lara, H. Boehnhardt, P.J. Gutiérrez . . . . . . . . . . . . . . . . . . . . . . . . . . . 287 Heliospheric energetic particle variability over the solar cycle D. Lario . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 295 Two years of Saturn’s exploration by the Cassini spacecraft: atmospheric studies A. Sánchez-Lavega, R. Hueso, S. Pérez-Hoyos . . . . . . . . . . . . . . . . . . . . . . . 303 A New Way for Exploring Solar and Stellar Magnetic Fields J. Trujillo Bueno . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 311 Session VII Observatories and instrumentation The MAGIC Telescopes (and beyond...) M. Martı́nez for the MAGIC Collaboration . . . . . . . . . . . . . . . . . . . . . . . . . . 321 Present and Future of Astronomy at the Observatorio del Teide A. Oscoz . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 331 Prospects for the William Herschel Telescope R. Rutten . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 339 VO Science. The Spanish Virtual Observatory E. Solano . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 345 Appendix: Table of Contents of the CD-Rom . . . . . . . . . . . . . . . . . . 353 Preface The Seventh Scientific Meeting of the Spanish Astronomical Society (Sociedad Española de Astronomı́a, SEA) was held at the Universitat de Barcelona in Catalonia from September 12 to 15, 2006. The event brought together 301 participants, who presented 161 contributed talks and 120 posters, the greatest number up to now. The fact that most exciting items of current astronomical research were addressed in the meeting proofs the good health of the SEA, a consolidated organization founded fifteen years ago in Barcelona. Two plenary sessions of the meeting were devoted to the approved entrance of Spain as a full member of the European Southern Observatory (ESO) and to the imminent first light of the largest telescope in the world, the GTC (Gran Telescopio de Canarias), milestones that will certainly lead the Spanish Astronomy in the near future. During the meeting, the SEA made public the Fourth Prize to the Best Spanish Ph. D. Thesis in Astronomy and Astrophysics for the period 2004-2005 ex aequo to Dr. Andrés Asensio Ramos and Dr. Patricia Sánchez Blázquez. The excellent PhD thesis of all applicants confirms the high level scientific career of young Spanish astronomers. Several specialists were invited to review central aspects in the different domains of astrophysics. We want to thank here all of them for their superb contributions. The effort of the Scientific Organizing Committee was crucial for the achievement of the scientific success of the meeting. The Society is indebted to the Universitat de Barcelona for hosting the meeting in its historical building; its pleasant and relaxing atmosphere sheltered and protected us against an intense Mediterranean storm. The Local Organizing Committee took care of all the logistic details to ensure a nice stay to all the participants. The meeting was possible thanks to the financial support of several research centers, universities, governmental institutions and private Spanish companies, these latest sharing with the astronomers the responsibility of achieving a great technological challenge in both, space and ground-based instrumentation. The proceedings of the Third, Fourth and Fifth Scientific Meetings of the SEA were published in the series “Highlights of Spanish Astrophysics”. Con- X Preface tributions to the Sixth SEA meeting, held in Granada in 2004 as a Joint European and National Astronomy Meeting, were published in the series JENAM Astrophysics Reviews. With this volume we return to the series “Highlights of Spanish Astrophysics”, with the hope of providing, now and in the future, a good compilation of the state of the art Spanish Astronomy to the international community. Invited talks are all collected in this book whereas contributed talks and posters presented at the meeting are published in the attached CD. Barcelona, January 2007 José Miguel Rodrı́guez Espinosa - SEA President Eduard Salvador Solé - LOC President Emilio J. Alfaro - SOC President Francesca Figueras Josep Miquel Girart Margarita Hernanz Carme Jordi Editors Scientific Organizing Committee Emilio J. Alfaro (president, Instituto de Astrofı́sica de Andalucı́a) Francesca Figueras (Universitat de Barcelona, IEEC) Francisco Garzón (Instituto de Astrofı́sica de Canarias) José Ignacio González (IFCA) Martı́n Antonio Guerrero (Instituto de Astrofı́sica de Andalucı́a) Margarita Hernanz (Institut de Ciències de l’Espai, CSIC-IEEC) Vicent J. Martı́nez (Observatorio Astronómico Universidad de Valencia) Josep Maria Paredes (Universitat de Barcelona, IEEC) Local Organizing Committee Marı́a Teresa Beltrán (Universitat de Barcelona, IEEC) Francesca Figueras (Universitat de Barcelona, IEEC) Josep Miquel Girart (Institut de Ciències de l’Espai, CSIC-IEEC) Guillermo González (Universitat Politècnica de Catalunya) Margarita Hernanz (Institut de Ciències de l’Espai, CSIC-IEEC) Carme Jordi (Universitat de Barcelona, IEEC) Rosario López (Universitat de Barcelona, IEEC) Belén López-Martı́ (Universitat de Barcelona, IEEC) Xavier Luri (Universitat de Barcelona, IEEC) Alberto Manrique (Universitat de Barcelona, IEEC) Josep Maria Paredes (Universitat de Barcelona, IEEC) Salvador Ribas (Universitat de Barcelona, IEEC) Ferran Sala (Universitat de Barcelona, IEEC) Eduard Salvador-Solé (president, Universitat de Barcelona, IEEC) Jordi Torra (Universitat de Barcelona, IEEC) José Marı́a Torrelles (Institut de Ciències de l’Espai, CSIC-IEEC) Sponsored by Ministerio de Educación y Ciencia (MEC through Programa Nacional de Astronomı́a y Astrofı́sica) Universitat de Barcelona (UB) Generalitat de Catalunya (Direcció General de Recerca, DGR) XII SOC,LOC,Sponsors Consejo Superior de Investigaciones Cientı́ficas (CSIC) Instituto Nacional de Técnicas Aeroespaciales (INTA) Instituto de Astrofı́sica de Canarias (IAC) Facultat de Fı́sica, Universitat de Barcelona (UB) Instituto de Astrofı́sica de Andalucı́a (IAA) Institut d’Estudis Espacials de Catalunya (IEEC) Universitat Politècnica de Catalunya (UPC) Fundació Catalana per a la Recerca (FCR) Grupo GMV Ingenierı́a y Servicios Aeroespaciales S.A. (INSA) Barcelona Aeronàutica i de l’Espai (BAIE) Centre de Supercomputació de Catalunya (CESCA) NTE, S.A. List of participants Acosta Pulido, José Antonio, Instituto de Astrofı́sica de Canarias Agueda Costafreda, Neus, Universitat de Barcelona-IEEC Alfaro Navarro, Emilio, Instituto de Astrofı́sica de Andalucı́a (CSIC) Alonso Herrero, Almudena, IEM (CSIC) Alvarez Pastor, José Manuel, Institut de Ciències de l’Espai, CSIC-IEEC Alves, Joao, Centro Astronómico Hispano Alemán Andrade Baliño, Manuel, Universidade de Santiago de Compostela Anglada Escudé Guillem, Universitat de Barcelona-IEEC Antoja Castelltort, M. Teresa, Universitat de Barcelona-IEEC Antón Ruiz, Luı́s, Universitat de València Aran Sensat, Àngels, Universitat de Barcelona-IEEC Arcones Segovia, Almudena, Max Planck Institut für Astrophysik Arregui Uribe-Echevarrı́a, Íñigo, Universitat de les Illes Balears Artal Garcı́a, Héctor, Universidad Autónoma de Madrid Artigas Roig, Anna, Institut de Ciències de l’Espai, CSIC-IEEC Ascası́bar Sequeiros, Yago, Astrophysikalisches Institut Potsdam Ascaso Anglés, Begoña, Instituto de Astrofı́sica de Andalucı́a (CSIC) Asensio Ramos, Andrés, Instituto de Astrofı́sica de Canarias Bachiller Garcı́a, Rafael, Observatorio Astronómico Nacional Badenes Montoliu, Carles, Rutgers University (EEUU) Balaguer Núñez, Lola, Universitat de Barcelona-IEEC Balcells Comas, Marc, Instituto de Astrofı́sica de Canarias Ballester Mortes, Josep Lluı́s, Universitat de les Illes Balears Barcons Jáuregui, Xavier, IFCA (CSIC-UC) Barrado Izagirre, Naiara, Escuela Superior de Ingenierı́a de la EHU Barreiro Vilas, R. Belén, IFCA (CSIC-UC) Barro Calvo, Guillermo, Universidad Complutense de Madrid Bayo Arán, Amelia, LAEFF (INTA) Bazán Casado, Juan José, Universidad Autónoma de Madrid Bellot Rubio, Luı́s Ramón, Instituto de Astrofı́sica de Andalucı́a (CSIC) Beltrán Sorolla, Maite, Universitat de Barcelona-IEEC Benavı́dez, Paula Gabriela, Universitat d’Alacant Benı́tez Lozano, Narciso, Instituto de Astrofı́sica de Andalucı́a (CSIC) XIV List of participants Benjouali, Latifa, Universidad Autónoma de Madrid Berná Galiano, José Ángel, Universitat d’Alacant Bernard, Edouard, Instituto de Astrofı́sica de Canarias Bihain, Gabriel, Instituto de Astrofı́sica de Canarias Bordas Coma, Pol, Universitat de Barcelona-IEEC Bosch Ramon, Valentı́, Universitat de Barcelona-Max Planck Institut für Kernphysik Busquet Rico, Gemma, Universitat de Barcelona-IEEC Bussons Gordo, Javier, IFCA (CSIC-UC) Caballero Garcı́a, Marı́a Dolores, LAEFF (INTA) Caballero Hernández, José Antonio, Max-Planck-Institut für Astronomie Cabezón Gómez, Rubén Martı́n, Universitat Politècnica de Catalunya Cabré Albos, Anna, Institut de Ciències de l’Espai, CSIC-IEEC Cabrera Lavers, Antonio Luı́s, Instituto de Astrofı́sica de Canarias Campo Bagatı́n, Adriano, Universitat d’Alacant Cantó Domènech, José, Escola Politècnica Superior d’Alcoi (UPV) Carballo Fidalgo, Ruth, Universidad de Cantabria Carbonell Huguet, Marc, Universitat de les Illes Balears Cardaci, Mónica, Universidad Autónoma de Madrid Cardiel López, Nicolás, Universidad Complutense de Madrid Carrasco Licea, Esperanza, INAOE Carrasco Martı́nez, José Manuel, Universitat de Barcelona-IEEC Carrera Jiménez, Ricardo, Instituto de Astrofı́sica de Canarias Carricajo Marı́n, Icı́ar, Universidade da Coruña Casalta, Joan Manel, NTE, S.A. Casas Rodrı́guez, Ricard, Agrupació Astronòmica de Sabadell Castander Serentill, Francisco Javier, Institut de Ciències de l’Espai, CSIC-IEEC Castañeda, Héctor, Instituto de Astrofı́sica de Canarias Castillo Morales, África, Universidad Complutense de Madrid Castro Rodrı́guez, Nieves, Instituto de Astrofı́sica de Canarias Castro Rodrı́guez, Norberto, Instituto de Astrofı́sica de Canarias Castro Tirado, Alberto J., Instituto de Astrofı́sica de Andalucı́a (CSIC) Català Poch, M. Asunción, Universitat de Barcelona Catalán Ruiz, Sı́lvia, Institut de Ciències de l’Espai, CSIC-IEEC Ceballos Merino, Maite, IFCA (CSIC-UC) Cenarro Lagunas, Javier, Universidad Complutense de Madrid Cepa Nogué, Jordi, Instituto de Astrofı́sica de Canarias Collados Vera, Manuel, Instituto de Astrofı́sica de Canarias Colomé Ferrer, Josep, Institut d’Estudis Espacials de Catalunya Colomer Sanmartı́n, Francisco, Observatorio Astronómico Nacional Comerón, Sebastien, Universitat Barcelona Comerón Tejero, Fernando, ESO-Garching Cordero Carrión, Isabel, Universitat de València Corral Ramos, Amalia, IFCA (CSIC-UC) Costado Dios, Teresa, Instituto de Astrofı́sica de Canarias Crespo-Chacón, Inés, Universidad Complutense de Madrid List of participants XV De Castro Rubio, Elisa, Universidad Complutense de Madrid De León Cruz, Julia Marı́a, Instituto de Astrofı́sica de Canarias del Olmo Orozco, Ascensión, Instituto de Astrofı́sica de Andalucı́a (CSIC) del Toro Iniesta, José Carlos, Instituto de Astrofı́sica de Andalucı́a (CSIC) Delgado Donate, Eduardo, Instituto de Astrofı́sica de Canarias Diago Nebot, Pascual David, Universitat de València Dı́az Beltrán, Ángeles Isabel, Universidad Autónoma de Madrid Dı́az López, Cristina, Universidad Complutense de Madrid Dı́az Sánchez, Anastasio, Universidad Politécnica de Cartagena Diego Rodrı́guez, José M., IFCA (CSIC-UC) Dierickx, Philippe, ESO -Garching (Germany) Diez Merino, Laura, ICMM-(CSIC) Docobo Durántez, José Ángel, Universidade de Santiago de Compostela Domingo Garau, Albert, LAEFF (INTA) Domı́nguez Aguilera, Inma, Universidad de Granada Ebrero Carrero, Jacobo, IFCA (CSIC-UC) Eiroa, Carlos, Universidad Autónoma de Madrid Eikenberry, Steve, Florida University (EEUU) Esquej Alonso, Pilar, Max-Planck-Institut für extraterrestrische Physik Estalella Boadella, Robert, Universitat de Barcelona-IEEC Fabricius, Claus, Universitat de Barcelona-IEEC Fernández Barba, David, Consorci del Montsec, Universitat de Barcelona-IEEC Fernández Soto, Alberto, Universitat de València Ferrer Soria, Antonio, IFIC (CSIC-UV) Ferreras Páez, Ignacio, King’s College London (UK) Ferri, Carlo, Institut de Ciències de l’Espai, CSIC-IEEC Figueras Siñol, Francesca, Universitat de Barcelona-IEEC Firpo Curcoll, Roger, IFAE Fors Aldrich, Octavi, Universitat de Barcelona Forteza Ferrer, Pep, Universitat de les Illes Balears Galadı́-Enrı́quez, David, Instituto de Astrofı́sica de Andalucı́a (CSIC) Gallart Gallart, Carme, Instituto de Astrofı́sica de Canarias Gallego Maestro, Jesús, Universidad Complutense Madrid Gálvez Ortı́z, Marı́a Cruz, Universidad Complutense de Madrid Garcı́a Benito, Rubén, Universidad Autónoma de Madrid Garcı́a Garcı́a, Mı́riam, Instituto de Astrofı́sica de Canarias Garcı́a López, Ramón, Instituto de Astrofı́sica de Canarias Garcı́a Melendo, Enrique, Fundació Observatori Esteve Duran Garcı́a Rojas, Jorge, Instituto de Astrofı́sica de Canarias Garcı́a Vargas, Marı́a Luisa, Instituto de Astrofı́sica de Canarias Garzón, Francisco, Instituto de Astrofı́sica de Canarias Gavilán Bouzas, Marta, Universidad Autónoma de Madrid Gil de Paz, Armando, Universidad Complutense de Madrid Gil-Merino Rubio, Rodrigo, The University of Sydney (Australy) Girart Medina, Josep Miquel, Institut de Ciències de l’Espai, CSIC-IEEC XVI List of participants Goicoechea Santamarı́a, Luı́s Julián, Universidad de Cantabria Gómez Martı́n, Cynthia, University of Florida Gómez Roldán, Angel, Revista Astronomı́a Gómez Velarde, Gabriel, Instituto de Astrofı́sica de Canarias González Casado, Guillermo, Universitat Politècnica de Catalunya González Fernández, Carlos, Instituto de Astrofı́sica de Canarias González Pérez, Violeta, Institut de Ciències de l’Espai, CSIC-IEEC Gorgas Garcı́a, Javier, Universidad Complutense de Madrid Gorosabel Urkia, Javier, Instituto de Astrofı́sica de Andalucı́a (CSIC) Guerrero Roncel, Martı́n Antonio, Instituto de Astrofı́sica de Andalucı́a (CSIC) Guirado Puerta, José Carlos, Universitat de València Guzmán Llorente, Rafael, University of Florida (EEUU) Hägele, Guillermo, Universidad Autónoma de Madrid Hernán Obispo, Marı́a Magdalena, Universidad Complutense de Madrid Hernanz Carbó, Margarita, Institut de Ciències de l’Espai, CSIC-IEEC Herrero Davó, Artemio, Instituto de Astrofı́sica de Canarias Hildebrandt, Sergi, Instituto de Astrofı́sica de Canarias Hirschmann, Alina, IEEC -UPC Hoyos Fernández de Córdova, Carlos, Universidad Autónoma de Madrid Huertas Company, Marc, Observatorio de Parı́s Hueso Alonso, Ricardo, Universidad del Paı́s Vasco Ibáñez Cabanell, José M., Universitat de València Iglesias Groth, Susana, Instituto de Astrofı́sica de Canarias Isasi Parache, Yago, Universitat de Barcelona-IEEC Isern Vilaboy, Jordi, Institut de Ciències de l’Espai, CSIC-IEEC Izquierdo Gómez, Jaime, Universidad Complutense de Madrid Jiménez Reyes, Sebastián, Instituto de Astrofı́sica de Canarias Jiménez Serra, Izaskun, IEM (CSIC) Jordi Nebot, Carme, Universitat de Barcelona-IEEC Julbe López, Francesc, Universitat de Barcelona-IEEC Kehrig, Carolina, Instituto de Astrofı́sica de Andalucı́a (CSIC) Khomenko, Elena, Instituto de Astrofı́sica de Canarias Labiano Ortega, Álvaro, Kapteyn Astronomical Institute Lara López, Luı́sa Marı́a, Instituto de Astrofı́sica de Andalucı́a (CSIC) Lario Loyo, David, The Johns Hopkins University (EEUU) Licandro Goldaracena, Javier, ING-IAC Lisenfeld, Ute, Universidad de Granada Loiseau Lazarte, Nora, ESAC López Aguerri, José Alfonso, Instituto de Astrofı́sica de Canarias López Corredoira, Martı́n, Instituto de Astrofı́sica de Canarias López Hermoso, Rosario, Universitat de Barcelona-IEEC López Martı́, Belén, Universitat de Barcelona-IEEC López Moratalla, Teodoro, Real Instituto y Observatorio de la Armada López Sánchez, Ángel R., Instituto de Astrofı́sica de Canarias López Santiago, Javier, Osservatorio Astronomico di Palermo (Italy) List of participants XVII Luna Bennasar, Manuel, Universitat de les Illes Balears Luri Carrascoso, Xavier, Universitat de Barcelona-IEEC Maldonado Prado, Jesús, Universidad Complutense de Madrid Manchado Torres, Arturo, Instituto de Astrofı́sica de Canarias Manera Miret, Marc, Institut de Ciències de l’Espai, CSIC-IEEC Manrique Oliva, Alberto, Universitat de Barcelona-IEEC Manteiga Outeiro, Minia, Universidade da Coruña Marcaide Osoro, Jon, Universitat de València Marco Tobarra, Amparo, Universitat d’Alacant Mármol Queraltó, Esther, Universidad Complutense de Madrid Martı́ Puig, José Marı́a, Universitat de València Martı́ Ribas, Josep, Universidad de Jaén Martı́ Vidal, Iván, Universitat de València Martı́n Manjón, Mariluz, Universidad Autónoma de Madrid Martı́n Pintado, Jesús, DAMIR-IEM-CSIC Martı́nez, Manel, IFAE Barcelona Martı́nez Arnáiz, Raquel Mercedes, Universidad Complutense de Madrid Martı́nez Carballo, M. Angeles, Instituto de Astrofı́sica de Andalucı́a (CSIC) Martı́nez Garcı́a, Vicent J., Observatori Astronòmic, Universitat de València Martı́nez González, Marı́a Jesús, Instituto de Astrofı́sica de Canarias Martı́nez Núñez, Sı́lvia, Universitat de València Martı́nez Pillet, Valentı́n, Instituto de Astrofı́sica de Canarias Martı́nez Serrano, Francisco Jesús, Universidad Miguel Hernández Martı́nez Vaquero, Luı́s Alberto, Universidad Autónoma de Madrid Masana Fresno, Eduard, Universitat de Barcelona-IEEC Masqué Saumell, Josep Maria, Universitat de Barcelona-IEEC Mateos Ibáñez, Sı́lvia, University of Leicester Mirabel, Félix, ESO -Chile Miralda Escudé, Jordi, Institut d’Estudis Espacials de Catalunya Miralles Torres, Juan Antonio, Universitat d’Alacant Moldón Vara, Javier, Universitat de Barcelona-IEEC Mollá Lorente, Mercedes, CIEMAT Montes Gutiérrez, David, Universidad Complutense de Madrid Montesinos Comino, Benjamı́n, Instituto de Astrofı́sica de Andalucı́a (CSIC) /LAEFF(INTA) Mora Fernández, Alcione, Instituto de Astrofı́sica de Andalucı́a (CSIC) Morales Calderón, Marı́a, LAEFF (INTA) Morales Durán, Carmen, LAEFF (INTA) Morales Peralta, Juan Carlos, Institut d’Estudis Espacials de Catalunya Morata Chirivella, Oscar, LAEFF (INTA) Moreno Lupiáñez, Manuel, Universitat Politècnica de Catalunya Muñoz Lozano, José A., Universitat de València Muñoz Mateos, Juan Carlos, Universidad Complutense de Madrid Najarro de la Parra, Francisco, IEM (CSIC) Negueruela Dı́ez, Ignacio, Universtat d’Alacant XVIII List of participants Nieto Isabel, Delfina Isabel, Universitat Politècnica de Catalunya Oscoz Abad, Alejandro, Instituto de Astrofı́sica de Canarias Oñorbe Berni, José, Universidad Autónoma de Madrid Padilla Torres, Carmen Pilar, Instituto de Astrofı́sica de Canarias Palau Puigvert, Aina, Universitat de Barcelona-IEEC Panessa, Francesca, IFCA (CSIC-UC) Paredes Poy, Josep Maria, Universitat de Barcelona-IEEC Pascual Ramı́rez, Sergio, Universidad Complutense de Madrid Pedraz Marcos, Santos, Observatorio Calar Alto Pelló Descayre, Roser, Observatoire Midi-Pyrénées Peralta Calvillo, Javier, Escuela Superior de Ingenieros Perea Duarte, Jaime, Instituto de Astrofı́sica de Andalucı́a (CSIC) Pérez Fournon, Ismael, Instituto de Astrofı́sica de Canarias Pérez Garcı́a, Ana Marı́a, Instituto de Astrofı́sica de Canarias Pérez González, Pablo G., University of Arizona (EEUU) Pérez Hoyos, Santiago, Universidad del Paı́s Vasco Pérez Martı́nez, Ricardo Manuel, ESAC Pérez Montero, Enrique, Universidad Autónoma de Madrid Pérez Torres, Miguel Ángel, Instituto de Astrofı́sica de Andalucı́a (CSIC) Perucho Pla, Manuel, Max-Planck-Institut für Radioastronomie (Germany) Pollock, Andrew, ESAC Portell Mora, Jordi, Universitat de Barcelona-IEEC Prada Martı́nez, Francisco, Instituto de Astrofı́sica de Andalucı́a (CSIC) Prieto Muñoz, Mercedes, Instituto de Astrofı́sica de Canarias Quilis Quilis, Vicent, Universitat de València Ramos Almeida, Cristina, Instituto de Astrofı́sica de Canarias Rebolo López, Rafael, Instituto de Astrofı́sica de Canarias Ribas Canudas, Ignasi, Institut de Ciències de l’Espai, CSIC-IEEC Ribas Rubio, Salvador José, Universitat de Barcelona-IEEC Ribó Gomis, Marc, CEA-Saclay, Universitat de Barcelona-IEEC Ribó Trujillo, Josep M., Universitat de Barcelona Riquelme Carbonell, Ma. Soledad, Universitat d’Alacant Rı́squez Óneca, Daniel, LAEFF (INTA) Roca Sogorb, Mar, Instituto de Astrofı́sica de Andalucı́a (CSIC) Rodes Roca, José Joaquı́n, Universitat d’Alacant Rodrı́guez Espinosa, José Miguel, Instituto de Astrofı́sica de Canarias Rodrı́guez Gasèn, Rosa, Universitat de Barcelona-IEEC Romero Gómez, Mercè, Universitat Rovira i Virgili Ros Ibarra, Eduardo, Max-Planck-Institut fuer Radioastronomie (Germany) Ruiz Camuñas, Ángel, IFCA (CSIC-UC) Rutten, Rene, Grupo de Telescopios Isaac Newton Sáez Milán, Diego Pascual, Universitat de València Sala Cladellas, Glòria, Max-Planck-Institut für extraterrestrische Physik (Germany) Sala Mirabet, Ferran, Universitat de Barcelona-IEEC Salvador Solé, Eduard, Universitat de Barcelona-IEEC List of participants Sanahuja Parera, Blai, Universitat de Barcelona-IEEC Sánchez Bejar, Vı́ctor Javier, GRANTECAN S.A. Sánchez Blázquez, Patricia, Laboratoire d’Astrophysique, EPFL Sánchez Gil, Ma. Carmen, Instituto de Astrofı́sica de Andalucı́a (CSIC) Sánchez Lavega, Agustı́n, Universidad Paı́s Vasco Sánchez Monge, Álvaro, Universitat de Barcelona-IEEC Sánchez Portal, Miguel, ESAC Sánchez Sánchez, Sebastián Francisco, Centro Astronómico Hispano Alemán Santander Vela, Juan de Dios, Instituto de Astrofı́sica de Andalucı́a (CSIC) Sanz Forcada, Jorge, LAEFF (INTA) Satorre Aznar, Miguel Ángel, Escola Politècnica Superior d’Alcoi (UPV) Serichol Augué, Núria, Universitat Politècnica de Catalunya Serra, Sinue, Universitat de Barcelona-IEEC Serra Ricart, Miquel, Instituto de Astrofı́sica de Canarias Sevilla González, Raúl, Universidad Autónoma de Madrid Sevilla Noarbe, Ignacio, CIEMAT Sidro Martı́n, Núria, IFAE Sierra González, Ma. del Mar, ESAC Solanes Majúa, José Marı́a, Universitat de Barcelona-IEEC Solano Márquez, Enrique, LAEFF (INTA) Soler Juan, Roberto, Universitat de les Illes Balears Suades Sol, Moisés, Institut de Ciències de l’Espai, CSIC-IEEC Telesco, Charles, University of Florida Toloba Jurado, Elisa, Universidad Complutense de Madrid Toribio Pérez, Ma. Carmen, Universitat de Barcelona-IEEC Torra Roca, Jordi, Universitat de Barcelona-IEEC Torrejón Vázquez, José Miguel, Universitat d’Alacant Torrelles Arnedo, José Marı́a, Institut de Ciències de l’Espai, CSIC-IEEC Torres, Diego F., Institut de Ciències de l’Espai, CSIC-IEEC Trancho Lemes, Gelys, Universidad de La Laguna/Gemini Observatory Trigo Rodrı́guez, Josep Maria, Institut de Ciències de l’Espai, CSIC-IEEC Trujillo Bueno, Javier, Instituto de Astrofı́sica de Canarias Ullán Nieto, Aurora, Universidad de Cantabria Vallbé Mumbrú, Marc, Instituto de Astrofı́sica de Canarias Vazdekis Vazdekis, Alexandre, Instituto de Astrofı́sica de Canarias Vicente Martı́nez, Ana Belén, Instituto de Astrofı́sica de Canarias Vilardell Sallés, Francesc, Universitat de Barcelona-IEEC Vı́lchez Medina, José Manuel, Instituto de Astrofı́sica de Andalucı́a (CSIC) Villamariz Cid, Charo, GRANTECAN S.A. Villar Pascual, Vı́ctor, Universidad Complutense de Madrid Willat, Rosemary, ESAC Yepes Alonso, Gustavo, Universidad Autónoma de Madrid Yun, Joao, Centro de Astronomia e Astrofı́sica da Universidade de Lisboa Zamorano Calvo, Jaime, Universidad Complutense de Madrid XIX XX List of participants Session I Spain in ESO Youth, accretion, and mass loss at the end of the main sequence F. Comerón ESO, Karl-Schwarzschild-Str. 2, D-85748 Garching bei München, Germany, fcomeron@eso.org Summary. The characterization of the properties of very low mass stars and substellar objects in star forming regions is an important research topic at ESO telescopes, and it has been pursued through the use of a wide variety of instruments and observing techniques. In this paper we focus on a project developed over the past few years devoted to the study of a particular group of very low mass, low luminosity objects in a number of nearby star forming regions that display strong indicators of accretion and mass loss. In this context, we also refer to related research carried out by other teams using ESO telescopes. Although the results amassed thus far do not allow us to unambiguously determine the nature of the objects exhibiting these characteristics, it appears that the examples studied until now span a wide range with regard to the relative importance of the accretion and mass loss signatures, the way in which the latter takes place, and possible also our vantage point with respect to them. This stresses the complexity of the earliest stages of stellar and substellar evolution, particularly regarding the comparison of evolutionary models that do not include mass accretion with the observational characteristics of real objects whose physical parameters are usually derived by using such models. 1 Introduction There is no doubt that we live in an era of renewed interest in the topic of circumstellar disks, mass loss, and their connection to the origin of stars. The reasons for this are manifold, but they may be grouped under three broad categories: a) the discovery of ever less and less massive substellar objects, which severely question basic assumptions of some models for the formation of stars and brown dwarfs; b) the connection between accretion and the formation of planetary systems, and the importance of dynamical interaction with a massive circumstellar disks leading to orbital migration as a determining factors in shaping the main features of the resulting planetary systems; and c) instrumental breakthroughs in sensitivity, resolution and spectral coverage achieved by unique facilities both on the ground and in space, which have enabled an increasingly direct observation of disks and jets. On the ground, such facilities 4 F. Comerón include adaptive optics and coronagraphic instrumentation nowadays available on 8m class telescopes, sensitive thermal infrared instruments that make it possible to image at the diffraction limit of such large apertures, and highresolution visible and near-infrared spectroscopy. Future unique facilities such as ALMA are ideally suited for high resolution imaging of cold dust and molecular line emission around stars and brown dwarfs. At the VLT interferometer (VLTI), upcoming improvements in fringe tracking and new instrumentation hold the promise of obtaining milliarcsecond-level spatial resolution even at faint magnitudes. From the space, the Spitzer observatory is living up to the expectations of revolutionizing the field by providing spectral energy distributions and spectroscopy of all but the outermost regions of disks around young objects at stellar and substellar masses, providing unprecedented insights into their structure and mineralogy, and there is no doubt that ESA’s Herschel observatory will follow on this path. 1.1 Some contributions at ESO Since this paper was delivered at the plenary session devoted to the accession of Spain to ESO, it is especially appropriate to outline here some of the major contributions that research groups making use of ESO’s facilities have made to the identification and study of the circumstellar environment of low mass stars. Such an enumeration is necessarily incomplete and subjective, but it hopefully serves the purpose of stressing the important role that state-of-theart instrumentation at ESO telescopes, most notably at the VLT, has played in the current knowledge in this field. It is also remarkable in the context of the present talk that several of the teams that have carried out this research are led by, or at least include, Spanish astronomers. • Deep imaging with narrow- and intermediate-band filters carried out with the Wide-Field Imager at the MPI/ESO 2.2m telescope on La Silla has detected faint, cool members of nearby Southern star forming regions thanks to the identification of Hα emission, providing statistically significant samples that demonstrate that Hα is still one of the preferred methods for the identification of young stellar objects, even at substellar masses ([1, 31, 32, 33]). • Thermal infrared imaging and spectroscopy of young stellar objects, with TIMMI2, ISAAC, and VISIR, have provided information on the largescale structure of disks around very low mass stars and massive brown dwarfs, as well as on the composition of their minerals and ices ([17, 41, 42, 47, 48, 50, 2, 3, 34]). • High-resolution spectroscopy in the visible carried out with UVES at the VLT has made possible the detailed analysis of the indicators of accretion and mass loss, as well as the monitoring on short timescales of their variability and rotational modulation. The application of innovative techniques such as spectroastrometry has also provided detailed in- Youth, accretion, and mass loss at the end of the main sequence • • • • • • 5 sights into the jet launch regions at the scale of a few astronomical units ([45, 46, 4, 26, 27, 51, 5, 39, 40, 52]). Radial velocity monitoring of very low mass objects in young star forming regions with UVES has investigated whether giant planets, particularly hot Jupiters, already form in the first few million years of their lives. Although no conclusive answer has yet been established, the results demonstrate that it is within the possibilities of current instrumentation ([24, 25, 49]). Characterization of young very low mass objects and their activity, using low resolution spectroscopy in the visible and near-infrared with VIMOS, FORS1/2, or ISAAC ([7, 8, 9, 20, 43, 23]). Spectroscopic characterization of disks and jets to the infrared, using SOFI and ISAAC, yielding emission-line diagnostics on deeply embedded sources ([39, 40, 37, 38]). Detection and study of sources surrounded by nearly edge-on disks, both with imaging (SOFI, ISAAC) and spectroscopy (UVES). Edge-on disk systems are particularly useful to study the circumstellar environment, as the disk acts as a natural coronagraph of the central source ([12, 22]). Detection of L0 -band infrared excesses with ISAAC, making it possible to identify very low mass members of embedded stellar aggregates and demonstrating that the excess emission due to warm circumstellar dust continues well beyond the substellar limit ([30]). Finally, we should mention the studies carried out by [11] using MIDI at the VLTI, which have provided unprecedented spatial resolution of inner disks around bright stars demostrating the power of mid-infrared interferometry. 2 Subluminous objects near the end of the main sequence In the rest of this paper we will focus on an intriguing class of young stellar objects that have received detailed attention in recent years due to the combination of apparently peculiar photospheric and circumstellar characteristics. Members of this class have spectra with intense emission lines, both those associated to outflows and to mass loss. The underlying photosphere has a late spectral type, and the BV RIJH colors are similar to those of normal stars, obscured by low to moderate extinction. K-band excess is weak or absent, and the very low luminosities, if taken at face value, place these objects on isochrones indicating ages much older than those expected in star forming regions, or even below the main sequence; see Figure 1. The accretion rates are estimated to lie in the 10−9 − 10−10 M /yr range. Despite these common features, important differences can be noticed among the different objects belonging in this class: on the one hand, the relative intensities of the lines used for accretion and mass loss diagnostics 6 F. Comerón Par−Lup3−4 ESO−Hα 569 ESO−Hα 574 LS−RCrA−1 Fig. 1. Temperature-luminosity diagram showing the positions of the four objects discussed in this paper with respect to evolutionary tracks ([6]) corresponding to typical ages in young aggregates. The small dots correspond to the stellar population at the center of the Chamaeleon I cloud ([14]), which are well matched by a 2 Myr isochrone. The horizontal locations of ESO-Hα 569 and 574 are only approximate, as their spectral types are poorly determined. change dramatically from object to object. On the other hand, their morphologies range from being unresolved even when observed at high spatial resolution, to resolved sources displaying hints of arcsecond-level structure, to the presence of well formed jets near the central source. Alhough we discuss in what follows four objects that we have discovered and followed up, earlier studies had already identified examples, such as HH 55, which was interpreted by [21] as a jet powered by a low-mass, relatively old member of the Lupus 3 cloud. Some of our objects also bear resemblance to 04158+2805, a possible class I source in the Taurus clouds ([29]). Other examples can be found in the literature (see [15]). It has been suggested (e.g. [10]) that the characteristics of these objects may be readily explained by edge-on disks blocking the light from a central source with the normal luminosity corresponding to the age of the aggregate to which it belongs. The blocking would account for the abnormally low luminosity, whereas the apparently normal colors would be a consequence of the photosphere of the object being seen mainly in scattered light. In turn, the prominence of the emission lines would be explained by the fact that they are produced at some distance from the central object, in regions directly visi- Youth, accretion, and mass loss at the end of the main sequence 7 ble. This may indeed explain some of these objects but, as we will see in the discussion on individual examples that follows, features observed in some of them may be difficult to account for in this scenario. 2.1 ESO-Hα 574: a low-luminosity object with a jet ESO-Hα 574 is a faint source in the periphery of the Chamaeleon I North cloud discovered by B. Reipurth in objective prism plates thanks to its Hα emission ([16]). Its spectral type is poorly determined, probably due to veiling of the photosphere by emission produced by accretion. Fig. 2. Spectrum of the central source of ESO-Hα 574, showing its rich emission line spectrum. The most prominent lines are marked. The underlying continuum is virtually featureless except for telluric absorption bands, possibly due to strong veiling. The emission-line spectrum is rich in forbidden lines (Figure 2), implying that mass loss signatures are dominant over those commonly attributed to accretion. Indeed, HeI emission is not detected and the intensity of the CaII triplet is very modest. Given the intensity of the emission lines, it is not surprising that deep narrow-band imaging in the [SII] filter ([13]) clearly shows a jet stemming from the central source (Figure 3). The bipolar jet is rather short, extending for only ∼ 3, 000 AUs of projected distance from end to end. Assuming that the emission-line spectrum at the position of the central source is similar to that along the jet, the derived physical parameters are similar to those of typical T Tauri jets. The central object is marginally resolved in near-infrared images, being elongated in a direction roughly perpendicular to the jet axis. The knotty structure of the northeastern jet suggests variability in the jet ejection parameters on timescales perhaps as short as one decade. 8 F. Comerón Fig. 3. Narrow-band image of ESO-Hα 574 obtained using FORS1 with a [SII] filter, clearly showing its the bipolar jet, HH 872. The resolved structure and its elongation, and the prominence of forbidden lines, lead us to consider ESO-Hα 574 as the object most likely to be an edgeon disk, as no characteristics known thus far conflict with this interpretation. 2.2 LS-RCrA 1 LS-RCrA 1, near the densest part of the R Coronae Australis star forming region, was discovered Hα emitter with 1.5 m Danish telescope on La Silla. Iits spectral type was determined shortly thereafter as M6.5 using FORS1 ([18]). Its location in the temperature-luminosity diagram corresponds to that of a 50 Myr old object. It displays prominent forbidden lines, HeI and CaII triplet emission, and weak H2 emission in the near-infrared. Colors are similar to those of a normal M6.5 star with light obscuration, with no K band excess, although showing hints of variability. The CO bands longwards of 2.29 µm ([18]) are significantly less deep than expected for an object of its spectral type. LS-RCrA 1 has been the subject of many follow-up observations by different groups, making it the most thoroughly studied object thus far in this class. Infrared observations under excellent seeing and adaptive optics imaging have failed to resolve it, placing rather stringent limits on any possible extended structure. Similarly, imaging with narrow-band filters in the visible and the near-infrared did not detect traces of a jet. Mid resolution spectroscopy has yielded further constraints on the nature of the emission-line spectrum. [10] confirmed a temperature close to that estimated from the spectral classification of [18] but inferred a surface gravity normal for a 8 Myr object, in better agreement with the age expectation for members of R CrA. They also reported single-peaked emission-line profiles, Youth, accretion, and mass loss at the end of the main sequence 9 Fig. 4. High-resolution UVES spectroscopy of forbidden lines in the spectrum of LS-RCrA 1. Note the clearly asymmetric profile of the [OI] lines, displaying a wing of blueshifted emission but no corresponding redshifted emission, probably due to the blocking of the red wing by a disk. The [NII] and [SII] lines, originating in regions of lower density and farther away from the central object, do not show such asymmetry, implying that the occultation of the emitting region occurs only in the close vicinity of the central object, where most of the [OI] emission originates. and concluded that their results supported the interpretation of LS-RCrA 1 being an edge-on disk system. Further arguments in this direction based on variability monitoring have been provided very recently by [44]. Nevertheless, new observations presented by [19] are difficult to reconcile with that scenario. Their higher resolution UVES spectra show broad wings of the Hα line with a width exceeding 300 km s−1 at 10 % intensity, which are thought to arise from the bases of accretion columns near the surface of the object ([35]) and are missing in the spectra of bona-fide edge-on systems, where that region is blocked from view ([4]). Most importantly, the [OI] lines are clearly asymmetric, with an extended blue wing but with the corresponding red wing missing, as shown in Figure 4. Such a profile is not seen in the [NII] or the [SII] lines, which are much more symmetric. Since the critical density of the [OI] emission is the highest among those species and this is thus the line forming closest to the star, we infer that the receding part of the inner outflow is occulted by a disk, thus excluding an edge-on geometry, whereas other forbidden lines predominantly forming further out are not affected by such blocking. Our UVES observations also show that the wings of the forbidden lines do not reach high velocities, and their profiles are clearly single-peaked, indicating that the outflow probably takes place in the form of a slow, weakly collimated wind, rather than in the form of a jet. 10 F. Comerón Recent VISIR and Spitzer data have provided the spectral energy distribution of LS-RCrA 1 up to the mid infrared, constraining the distribution of its circumstellar medium. Preliminary results (N. Huélamo, priv. comm.) indicate that no simple disk model can provide a satisfactory fit to the data, but also that the worst fits are obtained with a viewing geometry close to edge-on, seemingly in agreement with the conclusions reached from the spectroscopy. 2.3 Par-Lup3-4 Par-Lup3-4, discovered in a deep Hα slitless spectroscopy survey of the center of the Lupus 3 cloud carried out with FORS1 ([15]), has a spectrum similar to that of LS-RCrA 1, also with a late spectral type, M5, and a similar display of emission lines, although in this case the CaII triplet has more prominence. A comparison between the near-infrared photometry by [15] and that previously obtained by [36] shows that the object is clearly variable. There is no evidence for resolved emission at the position of the central source, although no images with the resolution of those available for LSRCrA 1 exist for Par-Lup3-4. However, narrow-band imaging through in the [SII] and Hα filters, also with FORS1 ([19]), clearly show an emission knot to the southwest 1”2 from the central source, corresponding approximately to 240 AU of projected distance. The knot is accompanied by a much fainter bipolar, well collimated jet in the northwest-southwest direction. UVES spectroscopy shows a double-peaked profile in the forbidden lines, particularly in [SII], with maxima separated by approximately 40 km s−1 . The latter demonstrates that the jet is not aligned with the plane of the sky, contrarily from the expectation if the features of the object were due to a perfectly edge-on disk. Assuming a spatial jet velocity in the range ∼ 100-150 km s−1 , typical of T Tauri stars, the tilt with respect to the plane of the sky is inferred to be between 8◦ and 12◦ , relatively close to edge-on. A recent examination of the spectral energy distribution using Spitzer data (F. Ménard, H. Bouy, N. Huélamo, priv.comm) independently confirms this value, obtaining a good overall fit for a tilt of 8◦ . Although the precise amount of obscuration caused by such a disk strongly depends on its vertical structure, it appears possible that Par-Lup3-4 may be intermediate between young stellar objects with an unobstructed view to the central object, and edge-on disks completely blocking the line of sight. The clearer line of sight towards the immediate circumstellar environment may help in explaining why Par-Lup3-4 displays broad wings in Hα, like LS-RCrA 1, together with strong CaII triplet lines and a well visible HeI line. Nevertheless, it remains to be demonstrated that the strong apparent underluminosity of Par-Lup3-4 is consistent with a lightly obscuring disk and its colors, particularly in the light of the fact that such a geometrical explanation is ruled out in the case of the otherwise similar LS-RCrA 1. Youth, accretion, and mass loss at the end of the main sequence 11 2.4 ESO-Hα 569 The last object in the sample that we have studied thus far, ESO-Hα 569, was discovered in the same objective prism survey of the Chamaeleon I cloud that led to the identification of ESO-Hα 574 ([16]). The features of both objects are to a first approximation similar: ESO-Hα 569 seems to possess a spectral type somewhat later than ESO-Hα 574, perhaps early M, although also with a large amount of veiling. The photometry of both objects is also similar, implying comparable amounts of underluminosity. However, when their emission-line spectra are compared, it becomes clear the ESO-Hα 569 and 574 are at opposite ends as far as the relative importance of accretion and outflow signposts are concerned, as seen in Figure 5. Indeed, ESO-Hα 569 displays the strongest lines of HeI and CaII measured among objects of this class, whereas the forbidden lines formed in jets or winds are by far the weakest when at all measurable. No traces of a well collimated outflow are noticeable in FORS1 [SII] images of ESO-Hα 569, but faint loop-like emission with low surface brightness towards the southwest can be seen near the detection limit of the available images. Unlike in the case of LS-RCrA 1, the central source of ESO-Hα 569 is clearly resolved in K-band images obtained under moderately good seeing, although the morphology is unclear. Fig. 5. A comparison between the spectra of ESO-Hα 569, an object dominated by lines commonly associated to accretion, and ESO-Hα 574, dominated by forbidden lines associated to mass loss. Both spectra define the extremes of the class of objects discussed in this paper. 12 F. Comerón 3 Preliminary conclusions and intriguing hypotheses While the four objects that we have discussed thus far share the essential characteristics of this class regarding apparent underluminosity, strong emission lines, and visible and near-infrared colors, it is somewhat surprising that their morphologies and probably also the inferred viewing geometries display such variety, leading one to wonder whether a single explanation may apply to all the objects under consideration and to others that may be related and that have been studied by other authors. The most immediate explanation for the underluminosity, an edge-on disk, may be the correct one of ESO-Hα 574 given the observations available to date. Partial obscuration by a disk may also apply in the case of Par-Lup3-4, but it seems unlikely in the case of LSRCrA 1 and probably also ESO-Hα 569 given the evidences already discussed regarding the occultation of the base of the outflow by a non-edge-on disk, the non-coincidence of the jet with the plane of the sky, or the high-velocity wings of the Hα line. We should remark that the discussion presented in this paper corresponds to a work in progress, and that new observations already scheduled for the coming months, particularly high resolution spectroscopy and imaging, may yield important new clues regarding the nature -or natures- of this class of objects. Some interesting possibilities can be already considered. If edge-on disks or, more generally, obscuration by the circumstellar environment can be confidently ruled out in at least some cases, the apparent underluminosity would then be a real, intrinsic characteristic of the central object. In previous works we have qualitatively invoked accretion-modified pre-main sequence evolution as a possible explanation for the observed features of these objects if truly intrinsic, based on modeling ([28]) at higher masses that in principle supported a false aging of accreting objects by increasing their temperature and decreasing their luminosity. However, the rather extreme two objects in Chamaeleon I in terms of underluminosity call into question such explanation, as accretion would then have to account for a decrease in radius to as little as 10 % of the value predicted by evolutionary models that do not take accretion into account. A rather exotic, and at this point rather speculative possibility is that the observed objects represent transient periods of intense accretion on much smaller objects, perhaps with masses in the giant planet range, temporarily increasing their temperatures and luminosities. If this were the case, the objects in this class, or at least those for which the low luminosity is not due to accretion, might be among the lowest-mass members of star forming regions, temporarily rendered visible due to the accretion bursts. It would be interesting in this respect to establish whether these objects have periods of quiescence, during which their spectral characteristics could be much more informative about the true nature of the central source. Regardless of whether such speculations or rather more mundane explanations should be invoked to account for the observed variety of phenomena in this class of objects, there is no doubt that the substellar boundary is a most Youth, accretion, and mass loss at the end of the main sequence 13 interesting place in terms of accretion, disk properties, and mass loss, and a promising and challenging territory for new instruments and facilities. Acknowledgements: I wish to thank Matilde Fernández as my main collaborator in this project, as well as other colleagues who have provided valuable insights on it. My thanks also go to Nuria Huélamo and Hervé Bouy for their interest and further work on LS-RCrA 1 and Par-Lup3-4. References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. Alcalá, J., Spezzi, L., Frasca, A. et al: A&A, 453, L1 (2006) Apai, D., Pascucci, I., Henning, Th. et al: ApJ, 573 (2002) Apai, D., Pascucci, I., Brandner, W. et al: A&A, 415, 671 (2004) Appenzeller, I., Bertout, C., Stahl et al: A&A, 434, 1005 (2005) Bacciotti, F., Eislöffel, J.: A&A, 342, 717 (1999) Baraffe, I., Chabrier, G., Allard et al: A&A, 337, 403 (1998) Barrado y Navascués, D., Zapatero Osorio, M.R. et al: A&A, 377, L9 (2001) Barrado y Navascués, D., Zapatero Osorio, M.R., Martı́n et al: A&A, 393, L85 (2002) Barrado y Navascués, D., Béjar, V.J.S., Mundt et al: A&A, 404, 171 (2003) Barrado y Navascués, D., Mohanty, S., Jayawardhana, R.: ApJ, 604, 284 (2004) van Boekel, R., Min, M., Leinert, Ch. et al: Nature, 432, 479 (2004) Brandner, W., Sheppard, S., Zinnecker, H. et al: A&A, 364, L13 (2000) Comerón, F., Reipurth, B.: A&A, 458, L21 (2006) Comerón, F., Neuhäuser, R., Kaas, A.A.: A&A, 359, 269 (2000) Comerón, F., Fernández, M., Baraffe, I. et al: A&A, 406, 1001 (2003) Comerón, F., Reipurth, B., Henry, A. et al: A&A, 417, 583 (2004) Dartois, E., d’Hendecourt, L., Thi, W.-F. et al: A&A, 394, 1057 (2002) Fernández, M., Comerón, F.: A&A, 380, 264 (2001) Fernández, M., Comerón, F.: A&A, 440, 1119 (2005) Gómez, M., Mardones, D.: AJ, 125, 2134 (2003) Graham, J.A., Chen, W.P.: AJ, 108, 2273 (2004) Grosso, N., Alves, J., Wood, K. et al: ApJ, 586, 296 (2003) Jayawardhana, R., Ivanov, V.D.: ApJ, 647, L167 (2006) Joergens, V., Guenther, E.: A&A, 379, L9 (2001) Joergens, V.: A&A, 448, 655 (2006) Hamilton, C.M., Herbst, W., Mundt, R. et al: ApJ, 591, L45 (2003) Hamilton, C.M., Herbst, W., Vrba, F.J. et al.: AJ, 130, 1896 (2004) Hartmann, L., Cassen, P., Kenyon, S.J.: ApJ, 475, 770 (1997) Kenyon, S.J., Brown, D.I., Tout, C.A. et al: AJ, 115, 2491 (1998) Lada, C.J., Muench, A.A., Lada, E.A. et al: AJ, 128, 1254 (2004) López-Martı́, B., Eislöffel, J., Scholz, A. et al: A&A, 416, 555 (2004) López-Martı́, B., Eislöffel, J., Mundt, R.: A&A, 440, 139 (2005) López-Martı́, B., Eislöffel, J., Mundt, R.: A&A, 444, 175 (2005) Meeus, G., Sterzik, M.F., Bouwman, J. et al: A&A, 409, L25 (2003) Muzerolle, J., Calvet, N., Hartmann, L.: ApJ, 550, 944 (2003) Nakajima, Y., Tamura, M., Oasa, Y.: AJ, 119, 873 (2000) Natta, A., Testi, L., Muzerolle, J. et al: A&A, 424, 603 (2004) 14 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. F. Comerón Natta, A., Testi, L., Randich, S.: A&A, 452, 245 (2006) Nisini, B., Bacciotti, F., Giannini, T. et al: A&A, 441, 159 (2005) Podio, L., Bacciotti, F., Nisini, B. et al: A&A, 456, 189 (2005) Pontoppidan, K.M., Dartois, E., van Dishoeck et al: A&A, 404, L17 (2003) Pontoppidan, K.M., Fraser, H.J., Dartois, E. et al: A&A, 408, 981 (2003) Saffe, C., Gómez, M., Randich, S. et al: A&A, 409, 993 (2003) Scholz, A., Jayawardhana, R.: ApJ, 638, 1056 (2006) Stempels, H.C., Piskunov, N.: A&A, 391, 595 (2002) Stempels, H.C., Piskunov, N.: A&A, 408, 693 (2003) Thi, W.-F., Pontoppidan, K.M., van Dishoeck, E.F. et al: A&A, 394, L27 (2002) Thi, W.-F., van Dishoeck, E.F., Dartois, E. et al: A&A, 449, 251 (2002) Torres, G., Guenther, E., Marschall, L.A. et al: AJ, 125, 825 (2003) van Broekhuizen, F.E., Pontoppidan, K.M., Fraser et al: A&A, 441, 249 (2005) Whelan, E.T., Ray, T.P., Bacciotti et al: Nature 435, 652 (2005) Ziener, R., Eislöffel, J.: A&A, 347, 565 (1999) The European Extremely large Telescope P. Dierickx European Southern Observatory pdierick@eso.org Summary. In 2000 the European Southern Observatory commissioned a conceptual study for a 100-m class optical telescope, dubbed OWL for its keen night vision. This study, undertaken with industrial and academic partners, was completed in November 2005 with the OWL concept design review. The panel of external experts concluded that the design was plausibly feasible and that the telescope could be operated. The OWL design, however, is evidently not without significant risks. In view of this, and of the expected cost (1.25 billion Euros), ESO Council decided to proceed towards an eventual European Extremely Large Telescope (E-ELT), with a diameter in the 30 to 60m range. Working Groups, with strong community participation, were set up to capture requirements and identify possible baselines. By May 2006, a range of options had been reviewed and two were retained for the design of a 42m telescope. Design and analysis work is now proceeding, with a view to selecting the baseline and entering the detailed design phase by early 2007, with construction starting by the end of 2009. In the following, we outline the design of OWL, its guiding principles, and summarize its strength and weaknesses. Thereafter we briefly explore design options for a 42m telescope and identify plausible design and trade-off directions. We also elaborate on ongoing technology developments, mostly the ELT Design Study, a generic technology development programme led by ESO and co-financed by the European Commission. 1 Introduction Building on the VLT experience and inspired by the success of optical segmentation (Keck telescopes), in the late 1990s ESO explored potential ways to build giant filled aperture telescopes, with a diameter of up to one hundred metre. Initial work concentrated on optical fabrication and feasibility of the structure, and prospective ideas were published in 1998 [5]. Following a positive response of industry, a design study for a 100-m adaptive, optical and near-infrared telescope, dubbed OWL for OverWhelmingly Large and for its keen night vision, was commissioned in 2000. Soon other projects emerged worldwide, with diameters ranging from 30 to 50 metre. Extremely Large Telescopes (ELTs) were nothing new; as early as 1977, Meinel et al. [7] had concluded that by that time a 25-m telescope was probably feasible, and since 1989 a group led by the University of Lund had been 16 P. Dierickx promoting a 25-m telescope concept [1], which eventually evolved into the 50m EURO50 [4]. The originality of OWL resided more in its aim at providing largest aperture at lowest cost rather than in its sheer size. To achieve that goal, it had to take a radically new approach towards design and fabrication, in particular reduce suppliers risks to the maximum possible extent and allow for compromises, as will be explained later on. In parallel with the OWL design study, the science case for extremely large telescopes was explored under the auspices of the Optical-Infrared Coordination Network (OPTICON), funded by the European Commission. A comprehensive report was released in 2005 [6] and work is still in progress, with strong support by the scientific community. 2005 also saw the start of the ELT Design Study, a generic scientific and technology development programme led by ESO and also funded by the European Commission within Framework Programme 6. With 25 partners in the industry and academia, the ELT Design Study aims at developing technologies and concepts crucial to any ELT, with little prejudice to actual size and design. Activities cover a broad range of topics, from wavefront control technologies to site characterization, enclosure concepts, to science requirements and integrated modelling, to name a few. The study is due for completion by 2008, with a good fraction of the deliverables becoming available in 2007. The OWL study was concluded in November 2005 [3, 8], and submitted to a panel of internationally recognized experts for review. In its executive summary, the panel concluded: “the team [has] demonstrated a plausible case that OWL is feasible and that a 100m telescope can be built and operated”. The panel also concluded that the scale and complexity of the project implied a high risk of schedule slippage, and that it would therefore not make a timely entry in the overall competitive scheme set by other extremely large telescopes and by the James Webb Space Telescope. The underlying technical arguments may be debated but the fundamental issue is not there. OWL represented too large a mind leap in too many areas: design, construction, operation, cost, and its science potential went way beyond what could be inferred incrementally from current knowledge. As a result, the support by the scientific community was, at best, hesitant, and supplier’s blessings were not sufficient to convince the community that the telescope was technically feasible. The panel nevertheless concluded on a positive note, recommending that ESO capitalizes on the OWL study and proceeds into the design phase of a smaller but still ambitious project. OWL became to mean Originally Was Larger until the acronym was (soon) dropped. In December 2005 topical working groups, with mixed community and ESO membership, had been set to provide community feedback, capture and prioritize requirements and establish the framework of a European ELT (EELT) in the 30- to 60-m range. The topics included science, telescope design, instrumentation, adaptive optics, and site aspects. After 3 months of extensive work, the working groups delivered their reports. An ELT Science and Engineering committee (ESE) was subsequently created, with non-ESO members The European Extremely large Telescope 17 only, to review the progress of the design and advise ESO on technical and scientific orientations of the project. In order to avert dilution of the effort into too many options, the baseline telescope diameter had been set to 42-m, i.e. midway between 30- and 60-m in terms of collecting area. At the lower end of the possible sizes, our colleagues in the US and Canada were developing a 30-m design, and at the higher end the OWL one could be adapted to 60-m. In June 2006 the E-ELT Project Office was officially created, and given the mandate to evaluate design options and propose a Basic Reference Design (BRD). The BRD would be presented to ESO Committees, to the scientific community, and eventually to ESO Council by the end of 2006, with a view to obtaining green light for detailed design. The plan calls for a detailed design phase over 2007-2009, a start of construction in 2010, first light by 2015, and full science operation by 2017. In a nutshell, the framework of the European ELT hinges on the following. First, a community-wide assessment of the science cases for an ELT (OPTICON); second, concept studies for 50- and 60 to 100-m telescopes (EURO-50, OWL); third, a broad technology development programme for enabling technologies (the ELT Design Study); fourth, the definition of a Basic Reference Design for a 42-m telescope; fifth, a robust European-wide academic and industrial expertise in critical areas (e.g. optical fabrication, instrumentation, adaptive optics). 2 The OWL concept The two highest priorities underlying the design of the OWL are low cost and low supply risks. Wherever possible, the design would rely on well proven technologies and industrially conscious solutions. For this reason classical designs with aspherical primary and secondary mirrors were quickly ruled out in favour of a 6-mirror design with spherical primary and flat secondary mirror. In addition to unmanageable costs, aspherical solutions also had significant system drawbacks, in particular a high sensitivity to decenters. A recurring concern is the effect of vibrations and wind in such large structure, and preference was given to designs minimizing the impact of decenters. The optical solution eventually selected after extensive trade-offs [2] is shown in Figure 1. It resembles the Southern African Large Telescope (SALT) solution and requires a four-mirror corrector with two active 8-m class mirrors (first stage, mainly compensating spherical aberration), a 4-m class passive and a 2.3-m flat adaptive mirror (second stage, compensating mainly field aberrations). The fourth mirror along the path of light has a very strong aspherization, but slope deviation from the best fitting sphere is comparable or lower to that of other systems already fabricated (SALT corrector) or being contemplated for other extremely large telescopes (e.g. the off-axis 8-m segments of the Giant Magellan Telescope). The primary and secondary mirrors are segmented, with 1.6-m segments flat-to-flat. The simple shape of the segments allows for a wide 18 P. Dierickx range of size and the final dimension is chosen for cost and compatibility with highly modular mirror cells. The spherical shape of the segments is compatible with large, stiff tool polishing i.e. suited for best optical quality. The telescope structure is optimized to favour shear deformation (lateral decenters) against tilt; the effect of primary-secondary mirror decenters is therefore limited and tight centering tolerances apply inside the corrector, which has a very stiff structure, instead of the entire telescope. Fig. 1. Layout of the OWL optical design. The telescope opto-mechanics has a modular (or fractal) design, with very high standardization i.e. it is made of nearly all-identical building blocks, each of which is composed of a limited number of different parts. The overall moving mass is about 14,800 tons, including contingency for paint, cabling, walkways, etc., i.e. very low for a structure of this size scaling the VLT up to 100-m would lead to about 60 times higher moving mass. At the same time it is reasonably stiff, with a 2.6 Hz locked rotor eigenfrequency. Static, dynamic and safety analysis show that the structure can be made of mild steel, with only minor reinforcements (higher grade steel) at specific locations to withstand earthquake loads. A preliminary dynamic analysis shows that the telescope could be operated in open air, assuming field stabilization with mirror M6 and with “soft” The European Extremely large Telescope 19 actuator technology for the segments active supports. A feed-forward control loop relying on accelerometers would allow to keep the phase of the segmented mirrors to a few nanometers with a wind speed of 10 m s−1 . Fig. 2. OWL telescope, overall layout. The opto-mechanical and control properties of the design are extensively described in [3, 8], and will not be recalled here. The overall characteristics are given in Table 1. The total cost estimate is 1.25 billion Euros (2005), including manpower, design, prototyping, capital nvestment and contingency. This figure includes industrial estimates for most of the capital investment (optics, mechanics, enclosure) and a supposedly generous allocation for adaptive optics (110 million Euros). Subsystems cost estimates derived from industrial studies, in particular for the primary and secondary mirrors, are remarkably consistent - most likely a consequence of relying on proven fabrication processes. The cost efficiency of OWL design is the result of a certain number of unique characteristics, most notably: • Modular design allowing for serial production (segments, structural modules, actuators, drives); • Low development and industrial risks for expensive items (e.g. spherical segments); • Open air operation allowing for a low-cost sliding enclosure; • No Nasmyth platform, allowing for maximum freedom in structural design (e.g. location of the altitude axis, balancing of the telescope). On the negative side, this design has a number of significant drawbacks: • A single mirror unit, M6, concentrates the most demanding wavefront control functions (field stabilization, adaptive optics); • Double segmentation (primary and secondary mirror), implying a complex wavefront control scheme; 20 P. Dierickx Table 1. OWL design, summary of characteristics. Entrance pupil diameter Focal ratio Total field of view Diffraction-limited field of view (Strehl Ratio≥0.80) λ = 0.5µm λ = 2.2µm λ = 5.0µm RMS spot size at edge of field (10 arc minutes) Central obscuration (linear) Emissivity (with pupil mask) Number of focal stations Primary mirror Secondary mirror M1-M2 separation M1 segments Number Optical shape Dimension (flat-to-flat) Thickness Substrate M2 segments Corrector Number Optical shape Dimension (flat-to-flat) Thickness Substrate M3 M4 M5 M6 Type Shape Radius of curvature Type Shape Diameter Radius of curvature Type Shape Radius of curvature Type Shape Tilt angle Diameter Number of actuators Control bandwidth Telescope mount Main structural material Main axes Drive and Bearing Systems Locked rotor frequency Gravity M1-M2 differential rigid body displacements Piston Tilt Decenter Tracking accuracy (Altitude and Azimuth axes only) Field stabilization range (M6 surface tip-tilt) Field Stabilization bandwidth Field Stabilization accuracy (M6 tip-tilt, before AO) 100-m 6.03 10 arc minutes 142 arc seconds (diameter) 245 arc seconds (diameter) 360 arc seconds (diameter) 0.052 arc seconds 35% 20.3% 6 Spherical, f/1.25 Flat, diameter 25.8-m 92517.5 mm 3048 Spherical, R=230-m 1.6-m 70 mm Low expansion glass / ceramic, Silicon Carbide as option 216 Flat 1.6-m 70 mm Low expansion glass / ceramic, Silicon Carbide as option Thin active meniscus Aspheric, concave, diameter 8250 mm 18690 mm Thin active meniscus Aspheric, concave; intermediate pupil 7800 mm 19970 mm Rigid mirror or thin adaptive shell Aspheric, concave, diameter 3950 mm 8504 mm Thin adaptive shell; exit pupil Flat 16o 2440×2660 mm2 98 across pupil 500 Hz Alt-az, rotating Mass 14834.5 tons Mild steel Friction Drive and Bearing 2.58 Hz 3.4 mm 13.1 arcsec 17.6 mm 0.3 arcsec rms with 10 m/s wind Min. ±31 arcsec PTV 2 Hz 0.01 arcsec rms The European Extremely large Telescope 21 • No gravity-stable instrument platform, inconvenient instrument location; • The enclosure provides little (if any) maintenance functionalities; • Hazard risks due to lengthy shutdown (enclosure would take about 30 minutes to cover the telescope). • At f/6, a short telescope focal ratio (arguably, this is a direct consequence of the diameter; a longer focal ratio would lead to an unreasonably large field of view at the telescope focus). In addition, the design delivers poor image quality of Laser Guide Stars (LGS), even after refocusing. Simulations having indicated that significant sky coverage could be obtained with Natural Guide Stars (NGS), the plan was to start operation with NGS only, and upgrade the telescope to LGS-capability with the 2nd generation adaptive systems. The review panel objected to this strategy; with the E-ELT, LGS priority has been raised. The OWL plan called for first light in 2016, start of science operations with a partially filled aperture in 2017 and a filled aperture by 2020. The 8-m mirrors of the corrector were on the critical path to first light, while segments integration rate was on the critical path to final completion. 3 The ELT Design Study In March 2004 a proposal for a technology development towards ELTs was submitted to the European Commission for funding within framework Programme 6. The proposal was approved, the project is running since January 1st, 2005, and is due for completion by end 2008, with major results available as early as 2007 in some areas. The project gathers 25 partners under ESO’s lead. The total estimated cost is M 29.4, including M 8.4 in Community support. ESO’s total contribution to the project amounts to M 11.740, out of which M 9.379 is covered by ESO internal funding. The participant’s own funding, by country of origin, is shown in Figure 3. The project covers the development of enabling technologies and concepts required for the construction of a European extremely large optical and infrared telescope. The ELT Design Study is complementary to the E-ELT design phase; both are conceived as parallel activities, the synergies and respective schedules allowing timely feedback between the two. The project breakdown includes 11 Work Packages, each of which is subdivided in several tasks. The technical Work Packages are outlined in Table 2. Detailed information about the project can be found in [8, 9]. Typical activities cover e.g. the development and on-sky testing of segments phasing techniques (Active Phasing Experiment or APE, Figure 4), the measurement of the performance of segments position control under representative excitation (Wind Evaluation Breadboard or WEB, Figure 5), breadboards (friction drives, Figure 6; adaptive mirror prototypes, etc.), instruments designs, site characterization, optical materials, and more. 22 P. Dierickx Table 2. ELT Design Study, technical work packages. Work Package Thematic tasks Science requirements Wavefront control Derive top level requirements from the science cases, and prioritize them Development and prototyping of metrology systems; delivery of position sensors for WEB. Development of segments support actuators, supply of 18 units for WEB. Characterization of diffractive properties, high contrast imaging studies and experiments. Active Phasing Experiment (APE). Wind Evaluation Breadboard (WEB). Optical Fabrication Production and testing of of 1-m class Silicon Carbide segment prototypes Mechanics Use of composite materials for targeted structural applications. Magnetic levitation as an alternative for the telescope kinematics. Breadboard friction drive. Enclosure Conceptual design of different types of enclosures; Computational Fluid Dynamics analysis, wind tunnel testing. Adaptive Optics Point designs, performance evaluation. Large deformable mirror technologies, prototyping. Novel AO concepts; LGS wavefront sensing, adaptable wavefront sensors. AO simulations, algorithms for reconstruction and control. Operations Operational models for an ELT Instrumentation Optical design of Atmospheric Dispersion Compensators. Point designs of instruments, detailed design of a representative subset. Site characterization Definition of site parameter space, standardization, design, supply and use of dedicated instrumentation. Characterization of atmospheric turbulence over large spatial scales. Integrated modeling Development of a generic integrated modelling tool. Conceived at a time where several possible telescope designs and dimensions were still being considered, the ELT Design Study was initially made independent of the actual telescope design. As designs converged in 2006 towards a single European project, minor adjustments were made to the scope of work to guarantee optimal synergy with the E-ELT project. 4 The European Extremely Large Telescope Following the OWL review, and with a view to moving from concept studies to a full-fledged project, topical working groups, with strong community representation, were established to capture requirements and lay the foundation of a 30- to 60-m European ELT. After three months intensive work, the working groups submitted their findings and recommendations in five areas: science requirements, telescope design, adaptive optics requirements and priorities, instrumentation, and site selection criteria. An ELT Science and Engineering (ESE) committee was subsequently created, with a view to advising the EELT project office as to project priorities and directions. The ESE is exclusively composed of non-ESO representatives and covers all topics addressed by the working groups. The telescope working group concentrated mostly on possible optical solutions. While there is a natural and to some extent healthy resistance to deviate The European Extremely large Telescope 23 Fig. 3. ELT Design Study, participants own funding, by country of origin. Fig. 4. Prototype segmented mirror for the Active Phasing Experiment (APE). from classical, well known solutions, the assumption that there should be a fit-for-all design is misplaced. A design is a response to requirements and constraints, and there is no priori reason for a solution to be equally valid for a 10-, a 42-, and a 100-m telescope. Spherical primary mirror solutions were dropped as the underlying cost benefit was thought to be marginal - a predictably unrealistic assumption, 24 P. Dierickx Fig. 5. Wind Evaluation Breadboard, Finite Element Model (courtesy IAC). Fig. 6. Friction drive (courtesy AMOS). Left: breadboard; right: drive unit. as shown by subsequent suppliers quotes. Practically, however, spherical primary mirror solutions would suffer from some of the weaknesses of the OWL design, and at 42-m, the cut between spherical and aspherical solutions is less clear. As of June 2006 work concentrated on two options: a Gregorian and a 5-mirror design (Figure 7). The latter is nothing else than a three-mirror anastigmat solution, with two flat relay mirrors. Both solutions rely on an f/1 aspherical primary mirror and provide a 10 arc minutes, f/15 field of view. The 5-mirror solution is diffraction-limited over the entire field and the field curvature is concentric to the exit pupil, a major advantage for instrumentation. The Gregorian has strong field curvature, convex towards the exit pupil. Similar structural designs have been produced and analyzed for both optical solutions. Structural performances are comparable. The telescope structure and the enclosure are inevitably larger with the Gregorian solution. Including at least one adaptive mirror in the telescope design is a strong advantage in many respects. First, it allows a single adaptive stage to feed The European Extremely large Telescope 25 Fig. 7. The European 42-m ELT; optical solution. several instruments; second, it simplifies the design of instruments-specific, specialized adaptive systems (e.g. extreme adaptive optics); third, the scale of the telescope and its exposure to natural conditions require faster control loop than with VLT-like active optics. In brief, the distinction between active and adaptive optics is a matter of bandwidth, spatial frequency and amplitude with existing 4- to 8-m telescopes; with a 42-m ELT the bandwidth and (owing to segmentation) spatial frequency range overlap to a significant extent; the distinction is essentially a matter of amplitudes. Drafting the error budget also shows that “classical” active optics would probably not deliver an optical quality comparable to that of the VLT. As a result, the E-ELT relies on a complex control scheme integrating active (first stage) and adaptive (second stage) optics (Figure 8). With the Gregorian, the secondary mirror is 4.6-m diameter, adaptive and fast steering for field stabilization. The mirror is segmented, with each segment supported by a kinematic mount. The Nasmyth relay flat mirror (tertiary mirror) is about 4-m in its largest dimension and probably needs to be segmented as well. This mirror is far from the pupil and its figure tolerances are unusually tight in order to limit its adverse (field-dependent) effect on adaptive corrections. With the 5-mirror solution, adaptive correction is provided by the 2.5m quaternary mirror and field stabilization by mirror unit M5, 2.7-m in its largest dimension. In the baseline, both mirrors are monolithic. The higher number of surfaces in the 5-mirror design leads to lower throughput and higher emissivity. It has been shown that this difference has 26 P. Dierickx Fig. 8. E-ELT wavefront control scheme (draft). negligible impact on overall performance (signal-to-noise, limiting magnitude), which is dominated by residual wavefront errors. Error budgets have been established. The first level breakdown includes residual atmospheric turbulence after adaptive correction, residual tip-tilt and decentering aberrations (effect of vibrations, wind buffeting), and intrinsic quality (telescope residuals in the absence of turbulence and external excitations). All three contributions are roughly similar in amplitude. The intrinsic quality is almost the same for both designs. The Gregorian has a lower number of surfaces but at least double segmentation, and the thin off-axis aspherical shells of the secondary mirror are inherently difficult to polish. In the 5-mirror design, the secondary is a 6-m VLT-like active mirror, the tertiary is a f/3 weak asphere, and mirrors M4 and M5 are flat. Atmospheric residuals are lower with the Gregorian, owing to its larger adaptive mirror hence higher number of actuators at constant interactuator pitch. For comparable AO residuals the 5- mirror requires, therefore, a densification of the actuator pattern. The European Extremely large Telescope 27 Residual tip-tilt under wind excitation is a major concern. A detailed analysis shows that telescope residuals impose demanding requirements on rejection at all frequencies, and that these requirements are significantly harder to meet with the Gregorian design: larger, heavier field stabilization mirror, inconvenient location (thin shell exposed to maximum wind speed). Fig. 9. The European 42-m ELT. The adaptive and field stabilization units of both designs have been the subject of three industrial studies. All studies concur that the Gregorian secondary is significantly more expensive and risky than the combined M3, M4, and M5 units of the 5-mirror design (one supplier declined to study the adaptive secondary mirror unit of the Gregorian). Expected delivery time is also 2-3 years longer. In conclusion, and taking into account all potential error sources, there does not seem to be a decisive performance advantage in favour of Gregorian, while there are certain risk, cost and schedule disadvantages. With limited (but still ambitious) extrapolation from present-day technology (a factor 2.5 in mirror diameter, a factor 5 in number of actuators), and assuming voicecoil actuators with 30 mm pitch (state of the art), the 5-mirror design should offer comparable performance with lower risk and cost, at an earlier stage. High actuator density is required mostly for small-field, short wavelength applications. An improvement (smaller pitch) or change in actuator technology (piezo instead of voice coils) would lead to better adaptive correction in the 5-mirror design. This could materialize in the baseline if allowed by a timely progress of the technology, as an upgrade if untimely. An interesting 28 P. Dierickx option would be to upgrade the tertiary mirror (also the least expensive unit) to a “second generation” adaptive one whence technology allows it. On grounds of cost, risk and schedule, the project office has selected the 5-mirror solution as baseline for its proposed Basic Reference Design. At the time of completion of this article, the Basic Reference Design has been presented to ESO Committees, to the scientific community, and to ESO Council. The response has been unequivocally positive, and ESO Council has decided to proceed with the detailed design phase. Acknowledgements: This report owes to the work of many scientists, engineers, technicians in the European academia and industry. There would be no talk about Extremely Large Telescopes without their enthusiastic support and contributions. Credit shall also be given to the European Commission for its support to the ELT Design Study and OPTICON (Framework Programme 6, contracts No 011863 and RII3-CT-2004-001566, respectively). References 1. Ardeberg, A., Andersen, T., Lindberg, B., Owner-Petersen, M., Korhonen, T., Søndergård, P., Breaking the 8m Barrier - One Approach for a 25m Class Optical Telescope, ESO Conf. And Workshop Proc. No 42, pp. 75-78 (1992) 2. P. Dierickx, B. Delabre, L. Noethe, OWL optical design, active optics and error budget; Proc.SPIE, 4003 (2000) 3. Dierickx, P., Gilmozzi, G., et al., OWL Concept Design Report, Phase A Review (2005) 4. Euro 50 book. 5. Gilmozzi, R., Delabre, B., Dierickx, P., Hubin, N., Koch, F., Monnet, G., Quattri, M., Rigaut, F., Wilson, R.N., The Future of Filled Aperture Telescopes: is a 100m Feasible?, Advanced Technology Optical/IR Telescopes VI, SPIE 3352, 778 (1998) 6. Hook, I. M. (Ed.), The Science Case for the European Extremely Large Telescope: The next step in mankind’s quest for the Universe (2005) 7. Meinel, A.B., An overview of the Technological Possibilities of Future Telescopes, ESO Conf. Proc. 23, 13 (1978) 8. http://www.eso.org/projects/owl/Phase− A− Review.html 9. http://www.eso.org/projects/elt-ds/ Gamma-ray bursts: lighthouses of the Universe J. Gorosabel Instituto de Astrofı́sica de Andalucı́a (CSIC), Apartado 3004, 18080 Granada, Spain, jgu@iaa.es Summary. In this paper some progresses in the field of gamma-ray bursts (GRBs) are briefly summarized. GRBs are the brightest explosions in the Universe, caused by the collimated ejection of ultrarelativistic matter from a powerful central engine and its subsequent collision with its environment. Due to their high luminosity GRBs are one of the most promising probes for cosmological studies. We discuss on the different classes of GRBs and their associated properties. The characteristics of the GRB host galaxies are also summarized. 1 Introduction As many other findings in science, GRBs were discovered serendipitously. In 1967-73, the four Vela satellites (named after the Spanish verb velar: to keep watch), that where originally designed for verifying whether the former Soviet Union abided by the Limited Nuclear Test Ban Treaty of 1963, observed 16 peculiarly violent high-energy events [62]. On the basis of arrival time differences, it was determined that they were not related neither to the Earth or the Sun, but they were of cosmic origin. A Gamma-Ray Burst (hereafter GRB) is an intense and brief pulse of γray radiation that occurs randomly on the sky. GRBs are not rare events in the Universe, occurring on our sky several times per day. They emit the bulk of their energy above ≈ 0.1 MeV, where our atmosphere is not transparent. So they have to be localized by detectors placed on board balloons, rockets or more usually on satellites. GRBs show a roomy morphological diversity of time profiles and a large range of durations, from a few milliseconds to several hundreds of seconds. By the mid eighties there was a general consensus that GRBs should be originated by Galactic neutron stars [73, 26]. However, in the nineties the BATSE instrument on board the Compton Gamma-Ray Observatory showed that the GRB sky distribution was highly isotropic with no concentration of events towards the Galactic plane [70]. This casted severe doubts on their Galactic origin. Also in the nineties it was discovered the 30 J. Gorosabel existence of two families of GRBs [63]: short (SGRBs, duration . 2 s) and long GRBs (LGRBs, duration & 2 s). Approximately 75% of the detected GRBs are LGRBs, being the rest SGRBs. In 1997, with the advent of the SAX satellite the first X-ray [18, 77], optical [93, 25, 74, 51, 10], infrared [15, 24, 42], millimeter [8] and radio [28, 36] counterparts were discovered. This allowed for the first time to measure the distance scale of GRBs [72, 10, 64], proving their cosmological origin. Thus GRBs became the most luminous sources in the Universe, releasing 1052−54 erg s−1 , assuming isotropic emission. This luminosity is comparable to burning up the entire mass-energy of the Sun in a few seconds, or to emit over that same period of time as much energy as our entire Milky Way does in a hundred years. Fig. 1 shows the absolute magnitude of two LGRBs in comparison to other Galactic and extragalactic sources as a function of time. As pictured LGRB afterglows are by far the optically brightest sources of the Universe if they are observed within 104−5 s with respect to the γ-ray event. -40 MV -30 -20 -10 0 1 10 Quasars Seyferts/Radio galaxies GRB 990123 GRB 021004 SN 1998bw SN Ia SN 1998S SN 1994I SN 1991bg Nova X-ray Binary 2 10 10 3 4 5 10 10 T-To (seconds) 6 10 10 7 8 10 Fig. 1. The V -band absolute magnitude of several Galactic and extragalactic sources. As displayed LGRB optical afterglows can reach Mv ∼ −35 in the first seconds. LGRBs are the most intense optical emitters, even more than Quasars, if they are detected with delays below 104−5 s with respect to the γ-ray emission. The discovered optical counterparts exhibited two main general properties: i) the fluxes faded approximately obeying power-law decays (Fν ∼ t−α ), and ii) the spectra were described by power laws (Fν ∼ ν −β ). These two findings were successfully described by the so-called afterglow model (see Sect. 2). Gamma-ray bursts: lighthouses of the Universe 31 A relevant discovery happened in 1999 with the observation of the extremely bright GRB 990123 [2]. The optical lightcurve of this GRB exhibited a clear deviation from i), showing a knee in the lightcurve which accelerated the flux decay ∼ 2 days after the GRB [11]. This break in the lightcurve is expected in a beamed geometry, when the Lorentz factor Γ of the outflow drops below θ−1 , being θ the jet opening angle. Around the lightcurve knee the observer starts to see the “edge” of the beam, so the observed flux is affected by a deficit of emitting material out of the jet boundaries. The temporal position of the break is dynamically tied to the θ value; the narrower the jet, the earlier the break time. Hence, measuring the time of the lightcurve knee it was possible to estimate the jet opening angle, and to correct the isotropic energy release. For the typical break values measured θ values around few degrees were inferred. This fact had deep energetic implications, decreasing the energy release by a Ω/4π ∼ θ2 /4 factor, and implying a luminosity reduction from 1052−54 erg s−1 to 1050−51 erg s−1 . At the same time it was noticed that not all GRBs show optical counterparts, albeit very deep and prompt searches were carried out [50, 43]. These gamma-ray bursts have been historically named as “Dark” GRBs (see [57] for a physical definition). At that time it was noted that for some GRBs, also called X-ray flashes (XRFs), the γ-ray photons were accompanied by a copious X-ray emission [53]. In 2005, more than 30 year after their discovery in γ-rays, the first optical counterpart to a SGRB was discovered [60, 78]. SGRBs show faint optical afterglows, being more elusive than LGRBs. To date the optical counterparts of only ∼10 SGRBs have been found, in contrast to LGRBs for which more than 100 optical afterglows have been localized since 1997. The paper is organized as follows: Sect.2 summarizes the afterglow model and Sect. 3 shows the central engines of LGRBs and SGRBs. Sect. 5 describes the GRB host galaxies, Sect. 4 explains their potential use for cosmological studies and Sect. 6 discusses a few future aspects. 2 The afterglow model The X-ray/optical/radio emission following a GRB can be modeled as the result of the shock of an ultrarelativistic ejecta with an external environment; the afterglow. In principle the afterglow model is valid for both LGRBs and SGRBs. This model does not explain the origin or central engine of the explosion, so it describes the observed properties once the gamma-ray emission is finished (several seconds after the GRB). It is interesting to note that the afterglow model was previous to the first afterglow discoveries in the nineties. The properties of the afterglow are explained when a compact source releases 1050−53 erg in a volume of ∼ 107−8 cm almost instantaneously. After ∼ 1 s, the internal energy has been transformed into kinetic energy because the plasma is optically thick and the fireball expands to ∼ 1010 cm. At some time, (may be with the central engine still active) around ∼ 101−2 s, the expanding 32 J. Gorosabel Fig. 2. The spectral energy distribution predicted for an isotropic afterglow model from radio to X-ray frequencies. Upper panel: the spectrum in the fast cooling regime, typically lasting few minutes, just after the GRB explosion. In this regime the expansion is modelized to follow two extreme cases; fully radiative (all the internal energy generated in the shock is radiated) or fully adiabatic expansion (the energy of the shock is constant). The three characteristic frequencies νa , νm and νc follow different fading laws depending whether the expansion is adiabatic (evolution laws over the arrows) or radiative (evolution law below the arrows, displayed in brackets). In both cases νc fades more rapidly than νm , so at some point νc catches νm . Lower panel: The bulk of the electron distribution (given by νm ) is below νc , so the radiative losses are negligible. Thus in the slow cooling regime the expansion is expected to be exclusively adiabatic. General: the above two regimes assume a spherical expansion, but still they are valid as long as the gamma factor Γ > θ−1 , where θ is the jet opening angle. Figure taken from [83] Gamma-ray bursts: lighthouses of the Universe 33 shell becomes optically thin and the photons can escape. The ultrarelavistic shell drives a forward blast wave (and likely a reverse one too) into the ambient medium, which sweeps up the interstellar matter. Assuming that the electrons are accelerated in the shock to a power-law distribution of Lorentz factor, it produces an afterglow at frequencies gradually declining from X-rays to visible and radio wavelengths. The pioneering afterglow models were improved including additional ingredients; calculations of radiative losses [83], diverse collimated outflow geometries [79, 84], reverse blast waves [71], ambient media with variable densities [16]. See [94] for a complete review. The afterglow spectrum can be described by four power-law segments defined by three characteristic break frequencies; νa , νc and νm (see Fig. 2). νa corresponds to the synchrotron self-absorbing frequency, below which the radio emission is self-absorbed and follows a blackbody shape (typically at a few GHz). The peak frequency νm is generated by the minimum-energy electrons and is associated to the maximum of the spectral energy distribution (SED). The cooling frequency νc corresponds to the high energy electrons which cool more rapidly than the characteristic expansion time. Above νa we find the standard low-frequency synchrotron slope F ∼ ν 1/3 up to either νm or νc , depending on the observing epoch. The decay of the three beak frequencies occurs at different rates, yielding two different emission regimes; fast and slow cooling (see Fig. 2). The transition between these regimes occur in few seconds/minutes, so follow up observations are typically performed in the slow cooling phase. The peak flux (Fm ) and the three break frequencies νa , νm , νc determine physical parameters as the ambient density, the isotropic energy release (assuming that the GRB redshift is known), and the fraction of the energy in electrons and magnetic fields [97]. It is interesting to note that as long as the outflow Γ factor is above θ−1 , the lightcurve/spectrum of an isotropic and a beamed afterglow are indistinguishable from the observer point of view. When Γ drops below θ−1 , then the evolution of Fm , νa , νm , and νc change with respect to the isotropic case [84]. An additional observational evidence of the afterglow beaming are the numerous detections of optical linear polarization [19, 96, 80, 20, 4, 21, 49, 69, 81, 44, 47] which are theoretically expected when Γ < θ−1 [41]. 3 The central engine The afterglow models assume an instantaneous release of energy in a reduced volume, but does not put any constraint on the source nature. What could produce such a violent energy injection? What is the central engine responsible of the GRB and the following afterglow emission? Do LGRBs and SGRBs share the same progenitors? 34 J. Gorosabel 3.1 The Collapsar model for LGRBs In the nineties hydrodynamic studies indicated that a collapse produced in a rapidly rotating massive star (the Collapsar) could release ∼ 1051 erg s−1 in γ-rays [98]. Fig. 3 displays the sequence of events occurring few seconds after the collapse. This hydrodynamic approach did a clear prediction; an underlying supernova (SN) should be significantly contributing to the optical flux measured ∼ 15 days after the GRB. The typical dynamical time of the whole hydrodynamic process is above a few seconds [67], so the Collapsar model found severe difficulties to explain the duration of SGRBs. Fig. 3. Schematic view of the Collapsar scenario. The collapse of a rapidly rotating massive star through an accretion disk (1) produces two bipolar jets that bore the star (2). Internal shocks in the jet generate the γ-ray responsible of the GRB (3). The interaction of the jet with the ambient medium produces the afterglow, responsible of the late emission from X-rays to radio wavelengths (4). The progenitor explodes as a SN reaching the optical lightcurve peak approximately in ∼15 days (5). In 1998 the bright Ic-type supernova SN 1998bw was found coincident with the long duration GRB 980425 [35], suggesting a link between some type of SNe and LGRBs, or at least with some of them. This first evidence was accumulating credence by detections in the optical lightcurves of “bumps” at ∼ 15 − 20 days after the GRB [6, 12, 13, 68, 45, 33, 5, 88, 85], as predicted by the Collapsar model. The final confirmation came from the intensive spectroscopic monitoring carried out for GRB 030329 around the time of the optical bump [54, 87]. As shown in Fig. 4 (right panel) the optical spectrum of the Gamma-ray bursts: lighthouses of the Universe 35 GRB 030329 afterglow suffered a metamorphosis and became a Ic-type SN (named as SN 2003dh) in a few days [54, 87]. Fig. 4. Left panel: The optical lightcurve of a typical afterglow. On a typical power law decay described by the afterglow model, an emission excess is detected due to an underlying supernova peaking ∼ 15 days after the GRB as predicted by the Collapsar model. Figure adapted from [45]. Right panel: spectral evolution of the GRB 030329 optical afterglow around the SN bump. The plot shows VLT spectra taken at several epochs, from 5 to 33 days after the GRB. As seen on April 3.10 (∼ 5 days after the GRB, upper line) the afterglow was dominated by synchrotron radiation and the spectrum was roughly a power law. In the following days the afterglow suffered a metamorphosis and became a typical Ic SN (bottom long dashed curve). Figure taken from [54]. Currently ∼ 10 GRB optical lightcurves have shown clear SN bumps. The fits of the optical lightcurves show that the commonly assumed template, SN 1998bw, does not universally reproduce the amplitude and the temporal width of the SN bump. In fact, the peak magnitudes of SNe associated to LGRBs show a dispersion of ∼ 1 magnitude [100]. It is interesting to note the intriguing cases of several SNe, like SN 2003lw, which showed high-energy emission similar to a GRB, but without any optical afterglow. SN 2003lw was accurately localized in X-rays and its position was angularly coincident with a bright z = 0.11 galaxy. Following the predictions of the Collapsar model, an intensive follow up of this galaxy, allowed for the first time the discovery of a SN spatially and temporally predicted ∼ 15 days earlier [91]. Also relevant were the observations of SN 2006aj associated to 36 J. Gorosabel the X-ray flash XRF 060218, for which the alert generated by the high-energy photons allowed the Swift satellite to observe a SN in real time. The lack of an optical afterglow for XRF 060218, allowed to record an extremely accurate SN lightcurve, since the first seconds of explosion [9, 76, 86]. Currently it is thought that there could be a continuous family of explosive events sharing the same physical mechanisms, which range from the least energetic SNe, XRFs, to the highly relativistic LGRBs. However, it is not properly understood why some high-energy events, apparently similar (all of them are related to stellar collapses), are able to (not) produce afterglows. It has been suggested that the mass and the angular momentum of the progenitor could determine the type of energetic event, and also the type of residual that the collapse might produce (neutron stars for XRFs, and black holes for LGRBs). Polarization measurements suggest that SNe related to XRFs/GRBs are highly asymmetric, resembling at some degree the GRB beamed geometry [76, 47]. See [99] for a review of Collapsar models. 3.2 Short duration GRBs: mergers of two compact objects The current knowledge on SGRBs is limited by the reduced well-studied SGRB sample. So the below discussion is still under debate. Observations carried out to date seem to indicate that SGRBs are less luminous than LGRBs approximately by an order of magnitude (∼ 1049 erg s−1 ) [55, 27]. Deep optical imaging did not detect any SN emission component in the lightcurve of the first SGRBs [56, 27, 14, 22], rejecting a connection of SGRBs with Collapsars. The most accepted models to explain the emission of γ-rays in a time interval as short as 10−3 s, are based on the coalescence of two compact objects (i.e. neutron star or black hole binary systems [1]). These collapsing systems show a geometry specially efficient to produce copiously gravitational waves. There is a fairly wide consensus that the emission of SGRBs is beamed [27], however still some aspects remain unsolved [95]. The main indirect arguments supporting the coalescence model are the following: i) the ambient medium density derived based on the SED is 3 orders of magnitude lower than that found in LGRB environments [82]. This low density supports the merger scenario, since the aged binaries should have time enough to travel in the host galaxy and tend to be located far away from their stellar birth places, where the density is much higher (as inferred for LGRBs, since they tend to occur in star forming regions). ii) A few angular measurements suggest that the positions of SGRBs show larger impact parameters than the ones measured in LGRB host galaxies [7]. This might indicate that at least a SGRB fraction is originated in the galaxy halos where a population of kicked-off neutron stars and black holes is expected. iii) The stellar population ages of SGRB host galaxies seem older[48] on average than the ones measured for LGRB hosts [17]. This is consistent with the coalescence model, given the long time needed to form the compact objects and to merge them by emission of gravitational waves. Gamma-ray bursts: lighthouses of the Universe 37 It has been noted that SGRBs show a lower mean redshift than LGRBs (∼ 0.6 vs. 2.8 [92, 59]). Furthermore, a positive correlation between BATSE error boxes of SGRBs and nearby bright galaxies was found, suggesting that a fraction (∼ 25%) of the detected SGRBs could occur in the local Universe (z < 0.025), in phenomena similar to the soft-gamma repeaters detected in the Magellanic clouds and caused by Magnetars [90]. However, it has been also suggested that a fraction of SGRBs could occur at high redshift and with energies comparable to LGRBs [92, 65, 3]. At high redshift the age of the Universe is not much higher than the coalescing time by emission of gravitational waves (108−9 s). Therefore, if high redshift SGRBs are originated by mergers, the coalescence models should be able to explain relatively short coalescence times (107−8 yr) by emission of gravitational waves. 3.3 Long duration GRBs with no SN: a new class of GRBs? Very recently it has been reported the discovery of 2 unusual GRBs [34, 23, 37] which might require a novel collapse scenario [38]. Examining the duration and properties of the γ-ray emission, it is accepted that these 2 GRBs should be classified as LGRBs. However, they do not show any SN brighter than MV ∼ −12.5 (100 times fainter than any SN ever observed) in clear contradiction with the Collapsar framework. Hence the traditional family separation between SGRBs and LGRBs will likely need a deep revision. 4 GRBs as cosmological rules Given the large redshifts at which GRBs occur it was evident their utility for cosmological studies. However, the isotropic GRB energies are not standard candles, so at a superficial glance GRBs do not seem suitable cosmic rules. Fortunately, as previously mentioned, the position of the optical lightcurve break is related to the beam opening angle [79, 84], so the beaming correction is possible. Once the beaming correction factor is introduced the γ-ray luminosity of GRBs is ∼ 1 dex wide, suggesting a quasi-standard energy reservoir [29]. The luminosity function was still too wide but an additional correction factor made finally GRBs competitive for cosmological studies: the energy peak of GRBs shows a correlation with the beaming-corrected energy release. This second correction factor allowed, using a sample of 15 GRBs with known redshift, to calibrate an empirical law that relates the peak energy of GRBs, with their γ-ray luminosity [40]. Therefore once a GRB is detected, a fit to the γ-ray spectrum yields the peak energy, and then from the empirical rule, it is possible to estimate the corresponding luminosity. This relation has provided a promising tool for cosmological tests. For instance, measuring the γ-ray flux on Earth, the luminosity (using the empirical relation) and the redshift (based on optical spectroscopic observations) the luminosity distance is fixed. The dependence of the luminosity distance with the cosmological parameters (Ωm , ΩΛ , H0 ) can be explored if a large enough GRB number is observed. 38 J. Gorosabel 5 Host galaxies of GRBs The redshift of LGRBs range from z = 0.0085[35] to z = 6.3 [52, 61, 89], so the corresponding hosts show a wide distribution of apparent magnitudes (15 < R < 29), centered around R ∼ 24.5. They tend to be blue and subluminous, exhibiting an intense star formation rate with respect to field galaxies at similar redshifts, specially if the specific star formation rate is compared [17]. The few metallicity measurements to date indicate submetallic values [46, 75]. Apart of a few exceptions, the morphology of LGRB hosts is irregular, in contrast to the core-collapse SNe hosts which tend to be brighter (very often grand-design spirals) [30]. LGRBs are powerful Lyman-α emitters [31, 32, 58]. SGRBs have been detected in both elliptical [39] and star forming galaxies [27], showing on average older stellar population ages than LGRB hosts [48]. Given the large angular separation seen in several SGRB afterglows, the association of a SGRB to its host galaxy is often based exclusively on statistical arguments, so it can be source of hosts’ miss-identifications [66]. 6 Future prospects The GRB discoveries carried out in the last 10 years has revealed a new multidisciplinary field, with natural links with neighbor disciplines, among many others, as studies of SNe, gravitational waves, astro-particles, or stellar evolution. The promising potential of GRBs for cosmological studies and to unveil the primitive universe, has created high expectation in the scientific community. The discovery of the first SGRB afterglow in 2005, has opened a new laboratory for crucial tests of theories of strong-field gravity, formation of black holes or to the nuclear equation of state. The detection of gravitational waves coincident to a SGRB by VIRGO, LIGO, TAMA, VIRGO or GEO experiments would imply a breakthrough similar to one occurred in 1974 with the General Relativity tests carried out using the binary pulsar PSR 1913+16. The future GLAST and AGILE missions, foreseen for 2007, will provide unprecedented sensitivity to gamma rays up to the GeV regime and detect ∼ 150 GRBs a year. These two missions will be complemented by EXIST, which will be sensitive to lower energies (10-600 keV). The combination of these spacecraft with rapid response ground-based facilities promises detailed spectroscopic/polarimetric multiwavelength campaigns. The role of the future “Gran Telescopio Canarias” (GTC) could be essential. The synergy or robotic telescopes with the GTC could allow to perform optical (+OSIRIS,+ELMER) and infrared (+EMIR) spectroscopic observations in the first minutes after the GRB. Hence using GRBs as lighthouses, the GTC could provide not only unprecedented information about the GRB itself, but also on all the intervening systems between us and the border of the optically known Universe. Gamma-ray bursts: lighthouses of the Universe References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. M.A. Aloy, et al.: A&A 436, 273 (2005) C. Akerlof, et al.: Nature 398, 400 (1999) E. Berger, et al.: ApJ submitted, astro-ph/0611128 (2006) D. Bersier, et al.: ApJ 583, L63 (2003) D. Bersier, et al.: ApJ 643, 284 (2006) J.S. Bloom, S.R. et al.: Nature 401, 453 (1999) J.S. Bloom, et al.: AJ 123, 1111 (2002) M. Bremer, et al.: A&A 332, L13 (1998) S. Campana, et al.: Nature 442, 1008 (2006) A.J. Castro-Tirado, et al.: Science 279, 1011 (1998) A.J. Castro-Tirado, et al.: Science 283, 1069 (1999) A.J. Castro-Tirado, J. Gorosabel: A&AS 138, 449 (1999) A.J. Castro-Tirado, et al.: A&A 370, 398 (2001) A.J. Castro-Tirado, et al.: A&A 439, L15 (2005) R. Chary, et al.: ApJ 498, L9 (1998) R.A. Chevalier, Zhi-Yun. Li: ApJ 536, 195 (2000) L. Christensen, et al.: A&A 425, 913 (2006) E. Costa, et al.: Nature 387, 783 (1997) S. Covino, et al.: A&A 351, 399 (1999) S. Covino, et al.: A&A 392, 865 (2002) S. Covino, et al.: A&A 400, L9 (2003) S. Covino, et al.: A&A 447, L5 (2006) M. Della Valle, et al.: Nature in press, astro-ph/0608322 (2006) A.H. Diercks, et al.: ApJ 503, L105 (1998) S.G. Djorgovski, et al.: Nature 387, 876 (1998) E.E. Fenimore, et al.: ApJ 335, L71 (1988) D.B. Fox, et al.: Nature 437, 845 (2005) D. Frail, et al.: Nature 389, 261 (1997) D. Frail, et al.: ApJ 562, L55 (2001) A.S. Fruchter, et al.: Nature 441, 463 (2005) J.P.U. Fynbo, et al.: A&A 374, 443 (2001) J.P.U. Fynbo, et al.: A&A 388, 425 (2002) J.P.U. Fynbo, et al.: ApJ 609, 962 (2004) J.P.U. Fynbo, et al.: Nature in press, astro-ph/0608313 (2006) T.J. Galama, M.S. et al.: Nature 398, 394 (1998) T. Galama, et al.: ApJ 500, L101 (1998) A. Gal-Yam, et al.: Nature in press, astro-ph/0608257 (2006) N. Gehrels, et al.: Nature in press, astro-ph/0610635 (2006) N. Gehrels, et al.: Nature 437, 851 (2005) G. Ghirlanda, et al.: ApJ 616, 331 (2004) G. Ghisellini et al.: MNRAS 309, L7 (1999) J. Gorosabel, et al.: A&A 335, L5 (1998) J. Gorosabel, et al.: A&A 339, 719 (1998) J. Gorosabel, et al.: A&A 422, L113 (2004) J. Gorosabel, et al. A&A 437, 411. (2005) J. Gorosabel, et al.: A&A 444, 711 (2005) J. Gorosabel, et al.: A&A 459, L33 (2006) 39 40 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. 88. 89. 90. 91. 92. 93. 94. 95. 96. 97. 98. 99. 100. J. Gorosabel J. Gorosabel, et al.: A&A 450, 87 (2006) J. Greiner, et al.: Nature 426, 157 (2003) P. Groot, et al.: ApJ 493, L27 (1998) A. Guarnieri, et al.: A&A 328, L13 (1997) J.B. Haislip, et al.: Nature 440, 181 (2006) J. Heise, J. et al.: 2003, in AIP Conf. Proc. 662, 229 J. Hjorth, et al.: Nature 423, 847 (2003) J. Hjorth, et al.: Nature 437. 859 (2005) J. Hjorth, et al.: 2005, ApJ 630, L117 (2005) P. Jakobsson, et al.: ApJ 617, L21 (2004) P. Jakobsson, et al.: MNRAS 362, 245 (2005) P. Jakobsson, et al.: A&A 447, 897 (2006) B.L. Jensen, et al.: GCN Circ 3589 (2005) N. Kawai, et al.: Nature 440, 184 (2006) R.W. Klebesadel, I.B. Strong, R.A. Olson: ApJ 182, L85 (1973) C. Kouveliotou, et al.: ApJ 413, L101 (1993) S.R. Kulkarni, et al.: Nature 393, 35 (1998) A. Levan, et al.: ApJ 648 , L9 (2006) A. Levan, et al.: ApJ submitted (2006) A.I. Mac Fadyen, S.E. Woosley: ApJ 524, 262 (1999) D. Malesani, et al.: ApJ 609, L5 (2004). N. Masetti, et al.: A&A 404, 465 (2003) C.A. Meegan, et al.: Nature 355, 143 (1992) P. Mészáros, M.J. Rees: MNRAS, 306, L39 (1999) M. Metzger, et al.: Nature 387. 878 (1997) T. Murakami, et al.: Nature 335, 234 (1981) H. Pedersen, et al.: ApJ 496, 311 (1997) B.E. Penprase, et al.: ApJ 646, 358 (2006) E. Pian, et al.: Nature 442, 1011 (2006) L. Piro, et al.: A&A 331, L41 (1998) P.A. Price, et al.: GCN Circ 2612 (2005) J. Rhoads: ApJ 525, 737 (1999) E. Rol, et al.: ApJ 544, 707 (2000) E. Rol, et al.: A&A 405, L23 (2003) P.W.A. Roming, et al.: ApJ 651, 985 (2006) R. Sari, T. Piran, R. Narayan: ApJ 497, L17 (1998) R. Sari, T. Piran, J.P. Halpern: ApJ 519, L17 (1999) A.M. Soderberg, et al.: ApJ 606, 994 (2004) J. Sollerman, et al.: A&A 454, 503 (2006) K.Z. Stanek, et al.: ApJ 591, L17 (2003) K.Z. Stanek, et al.: ApJ 626, L5 (2005) G.C. Tagliaferri, et al.: A&A 443, L1 (2006) N.R. Tanvir, R. Chapman, A.J. Levan, R.S Priddey: Nature 438, 991 (2005) B. Thomsen, et al.: A&A 419, L21 (2004) A. de Ugarte Postigo, et al.: ApJ 648, L83 (2006) J. van Paradijs, et al.: Nature 386, 686 (1997) J. van Paradijs, C. Kouveliotou, R.A.M.J. Wijers: ARA&A 38, 379 (2000) D. Watson, et al.: A&A 454, L123 (2006) R.A.M.J. Wijers, et al.: ApJ 523, L33 (1999) R.A.M.J. Wijers, T.J. Galama: ApJ 519, L17 (1999) S.E. Woosley: ApJ 405, 273 (1993) S.E. Woosley, J.S. Bloom: A&ARA 44, 507 (2006) A. Zeh, S. Klose, D.H. Hartman: ApJ 609, 95 (2004) The VIMOS VLT Deep Survey (VVDS) R. Pelló1 , O. Le Fèvre2 , C. Adami2 , M. Arnaboldi3 , S. Arnouts2 , S. Bardelli4 , M. Bolzonella4 , A. Bongiorno4, M. Bondi5 , D. Bottini6 , G. Busarello3 , A. Cappi4 , S. Charlot7 , P. Ciliegi4 , T. Contini1 , S. Foucaud6 , P. Franzetti6 , B. Garilli6 , I. Gavignaud8, L. Guzzo9 , O. Ilbert4 , A. Iovino9 , F. Lamareille1 , V. Le Brun2 , D. Maccagni6 , B. Marano4, C. Marinoni9 , G. Mathez1 , A. Mazure2 , H.J. McCracken7, Y. Mellier7 , B. Meneux2 , P. Merluzzi3 , R. Merighi4 , S. Paltani2 , J.P. Picat1 , A. Pollo9, L. Pozzetti4 , M. Radovich3 , V. Ripepi3 , D. Rizzo10 , R. Scaramella5, M. Scodeggio6 , L. Tresse2 , G. Vettolani5 , A. Zanichelli5 , G. Zamorani4, and E. Zucca4 1 2 3 4 5 6 7 8 9 10 Observatoire Midi-Pyrénées, Laboratoire d’Astrophysique, UMR 5572, 14 Avenue E. Belin, F-31400 Toulouse (France), roser@ast.obs-mip.fr Laboratoire d’Astrophysique de Marseille (UMR 6110), CNRS-Université de Provence, BP8, F-13376 Marseille Cedex 12 (France), Olivier.LeFevre@oamp.fr Osservatorio Astronomico di Capodimonte INAF (Italy) Osservatorio Astronomico di Bologna INAF (Italy) Istituto di Radio-Astronomia, INAF (Italy) IASF, INAF, 20133 Milano (Italy) Observatoire de Paris (France) European Southern Observatory (Germany) Osservatorio Astronomico di Brera INAF (Italy) Imperial College of Science, Technology and Medicine, London (UK) Summary. This paper1 reviews the main results obtained so far by the VIRMOS VLT Deep Survey (VVDS) collaboration. The VVDS is one of the leading extragalactic surveys, started in 2002 using GTO awarded for the construction of VIMOS-VLT. Its aim is to study the evolution of galaxies, large scale structures and AGNs from a goal sample of ∼100000 objects down to a magnitude I(AB)=24, with redshifts measured using VIMOS-VLT (typically in the range 0<z<5), and a wide wavelenght coverage from different follow-up observations. Among the most important results achived by the VVDS so far are the redshift distributions for different galaxy populations, the evolution of the galaxy Luminosity Function in various filter bands (as a function of redshift, galaxy type and environment), the evolution of the typedensity relation with redshfit, and the clustering properties from z∼2. One of the most impressive findings of the VVDS using a pure magnitude-selected sample is 1 Based on data obtained with the European Southern Observatory Very Large Telescope, Paranal, Chile, program 070.A-9007(A), and on data obtained at the Canada-France-Hawaii Telescope, operated by the CNRS of France, CNRC in Canada and the University of Hawaii. 42 Pelló et al. the existence of an important population of bright galaxies at 1.6<z<5, 1.6 to 6.2 times larger than previous estimates based on color pre-selections (LBG technique). 1 The VIMOS VLT Deep Survey: an overview The VIMOS VLT Deep Survey2 (hereafter VVDS) is a breakthrough spectroscopic survey which is intended to provide a unique view of the universe at 0≤z≤5, thanks to the impressive multiplexing capabilities of the VIMOS spectrograph. VIMOS 3 is a genuine “Redshift Machine” built for ESO-VLT by a Franco-Italian consortium. It is a wide field imager and multi- object spectrograph in the visible domain (0.36 to 1 µm), mounted on the Nasmyth focus B of UT3 Melipal. It has four identical arms, each one with a field of view of 7’ x 8’, and it provides a spectral resolution in the range ∼200-2500. Up to ∼1000 spectra can be accomodated in a single shot in the low-resolution mode. More information about the capabilites of VIMOS can be found in the reference paper by [6]. The VVDS has been carried out using mainly the consortium’s nights of guaranteed time. About 50 scientists in seven different institutes in France and Italy have participated to this survey. In Sect. 2 we present an overview of the survey design and strategy, together with the current status. We also review in this section the main results obtained so far by the VVDS on the formation and evolution of galaxies: redshift distribution for different magnitude-selected samples; evolution of the galaxy Luminosity Function (LF) as a function of redshift and local density for different spectrophotometric types of galaxies; the build-up of the colour-density relation with cosmic time; the evolution of clustering; the properties of a unique sample of faint Type 1 AGNs, and the properties of the high-reshift galaxy population at 1.5≤z≤5. Sect. 3 summarizes the results and conclusions. The cosmological parameters adopted throughout this paper are Ωm = 0.3, ΩΛ = 0.7, and H0 = 70 km s−1 Mpc−1 . 2 Constraining the formation and evolution of galaxies 2.1 Survey Design and Current Status The VVDS includes two main steps, the imaging survey and the spectroscopic follow up. Deep UBVRI imaging on the four 2×2 deg2 fields listed in Table 1, as well as K’ imaging over selected areas, was obtained with the CFHT-CFH12K camera (BVRI), the ESO-NTT-SOFI (K’), and the ESO2.2m-WFI (U). Data reduction and processing was carried out at Terapix4 . Photometric catalogs were obtained with SExtractor [1]. The Shallow Survey 2 3 4 http://www.oamp.fr/virmos/vvds.htm http://www.eso.org/instruments/vimos/ http://terapix.iap.fr The VIMOS VLT Deep Survey (VVDS) 43 includes the four 2 × 2 deg2 (16 deg2 in total), with typical depth of AB∼24.5 to 26.0, depending on filters. The Deep Survey includes one pointing of 2 deg2 on 0226-0430 up to AB∼26 to 27 [7]. There are three different parts in the VVDS spectroscopic follow up: the “Wide” Survey (IAB ≤22.5), the “Deep” Survey (IAB ≤24.0) and the “UltraDeep” (IAB ≤24.75). Part of the deep survey was obtained in an area of 21 × 21.6 arcmin2 around the Chandra Deep Field South (CDFS) [8]. The current status of redshift measurements in the VVDS are summarized in Table 1. The present data cover about 1.2 deg2 in the Deep Survey and 10 deg2 in the Wide Survey. The goal is to complete a final sample of ∼100000 spectra on ∼10 deg2 . Table 1. Current status of redshifts measured in the VVDS Field VVDS-0226-04 VVDS-1000+03 VVDS-1400+05 VVDS-2217+00 CDFS TOTAL IAB ≤22.5 IAB ≤24.0 IAB ≤24.75 ∼14000 (public end 2006) 1000 (on-going) ∼5000 ∼11000 ∼15000 ∼10000 (on-going) ∼1600 (public) ∼35000 ∼15600 ∼1000 2.2 Measuring galaxy evolution from the VVDS The VVDS efficiently covers the 0.0≤z≤5 domain. It is successfully going through the redshift desert at 1.5≤z≤2.2, while the 2.2≤z≤2.7 domain remains of difficult access due to the wavelength coverage. The VVDS reaches a completeness level in redshift measurements of 78% overall (93% including less reliable flag 1 objects), with a typical spatial sampling of the galaxy population ranging between ∼25% and ∼30% in the Deep sample (VVDS-02h and VVDS-CDFS). Deep iAB ≤ 24 redshift distribution The redshift distribution N(z) obtained in the Deep sample has a median of z=0.62, z=0.65, z=0.70, and z=0.76, for magnitude-limited samples with IAB ≤ 22.5, 23.0, 23.5 and 24.0 respectively. A high redshift tail above redshift 2 and up to redshift 5 becomes readily apparent for IAB ≥ 23.5, probing the bright star-forming population of galaxies (see below and [9]). 44 Pelló et al. Galaxy Luminosity Function The evolution of the galaxy LF in different rest-frame bands has been studied with the VVDS data in different ways (e.g. [4], [5] and [15]). Considering the global LF, a substantial evolution is observed as a function of redshift in all bands. Up to z = 2, a brightening of the characteristic magnitude M∗ is observed (Fig. 1). The comoving density of bright galaxies (i.e. brighter than M∗ (z=0.1)) increases by a factor of ∼2.6, 2.2, 1.8, 1.5, 1.5, between z=0.05 and z=1, in the U, B, V, R, I bands, respectively. A possible steepening of the faint-end slope of the LF is also measured, with ∆α = -0.3 between z=0.05 and z=1, similar in all bands. When considering the rest-frame Bband LF for two different populations of bulge and disk-dominated galaxies at z=0.4-0.8, the LF slope is found to be significantly steeper for the diskdominated population (α = −1.19 ± 0.07) compared to the bulge-dominated population (α = −0.53 ± 0.13). The red bulge-dominated galaxies are already in place at z∼1, but the volume density of this population increasses by a factor 2.7 between z∼1 and z∼0.6. These observations are consistent with the faint and compact population of galaxies as possible progenitors of the local dwarf spheroidal galaxies ([15]). When computing the LF as a function of the local density of galaxies, the slope is found to be systematically steeper in under-dense environments, whereas there is no significant difference in M∗ between over- and under-dense environments ([5], Fig. 1). Another interesting result is the LF for galaxy samples selected by spectral type out to z=1.5 ([5]), in several rest-frame bands. The VVDS sample allows this exercice for over 70% of the age of the Universe. Galaxies have been classified in four spectral types, from early type to irregular galaxies, using their colours and redshift. Luminosity functions have been computed in the U, B, V, R and I rest frame bands for each type, in redshift bins from z=0.05 to z=1.5. Some illustrative examples are given in Fig. 1. A significant steepening of the LF is found from early to late types. M∗ is significantly fainter for late type galaxies, and the difference between types increases in the redder bands. A brightening of M∗ is found for all spectral types with increasing redshift, ranging from ∼0.5 mag for early type galaxies to ∼1 mag for the latest type galaxies. On the contrary, the slope of the LF is consistent with no-evolution with redshift for all spectral types. For early type galaxies, the LF is consistent with passive evolution up to z ∼1.1, whereas the number of bright (MB (AB)≤ -20) galaxies decreases by ∼40% from z∼0.3 to 1.1. For the latest type galaxies, the normalisation of the LF strongly evolves with redshift, with an increase by more than a factor of 2 between z∼0.3 and 1.3, in such a way that the density of bright galaxies increases of a factor ∼6.6 in this redshift domain. The build-up of the colour-density relation with cosmic time The study of the redshift and luminosity evolution of the galaxy colour-density relation in the VVDS has been recently published by Cucciati et al. ([2]). In- The VIMOS VLT Deep Survey (VVDS) 45 E/S0 over−dense over−dense under−dense under−dense Im Fig. 1. Left: Evolution of the LF in the B-band for two extreme spectrophotometric types of galaxies (top:E/S0; bottom: Im). Right: Evolution of the B-band LF as a function of the local density (adapted from Ilbert et al. 2006). deed, the VVDS sample allows to reconstruct the 3D environment of galaxies on R=5h−1 Mpc scales up to z∼1.5. The colour-density relation exhibits dramatic changes as a function of redshift. A steep colour-density relation is confirmed at low redshifts, with the fraction of the reddest(/bluest) galaxies of the same luminosity increasing(/decreasing) as a function of density. This trend progressively disappears towards the highest redshift bins. The rest frame u∗ -g’ colour-magnitude diagram shows a bimodal distribution in both low and high density environments up to z∼1.5. The bimodality is not universal but strongly environment and luminosity dependent. The star formation activity is found to be progressively shifting with decreasing redshift towards lower luminosity galaxies in low density environments. Evolution of Clustering from z∼ 2 Clustering is one of the most important dignostics of galaxy evolution through cosmic times. The VVDS sample is particularly well suited to probe sufficiently large volumes and separation scales at different epochs, for different galaxy types. Different results have been obtained on clustering properties, 46 Pelló et al. covering the redshift domain 0≤z≤2 [10] [12] [14]. The global correlation length is found to be 3.6h−1 Mpc for a magnitude-limited sample with IAB ≤24.0 at 1.3≤z≤2.1 (M∗ ∼-21). Compared to galaxies in the local universe, this result indicates that clustering has strengthened with cosmic time, with the correlation amplitude increasing by ∼2.5 times to provide the local values for galaxies with similar luminosities [10]. When considering the clustering of galaxies as a function of the spectroscopic type up to z∼1.2, early-type galaxies are found to be more strongly clustered than late types [12]. On the other hand, at z∼1 the most luminous galaxies are more strongly clustered than the fainter ones [14]. This is an indication that, at these redshifts, the most evolved galaxies were already sitting in the stronger density peaks. Faint Type 1 AGNs Type 1 active galactic nuclei (AGN) have been extracted from the VVDS magnitude-limited sample: 130 broad-line AGN spectra (BLAGN) up to z∼5, divided into wide (IAB ≤22.5) and deep (IAB ≤24.0) subsamples containing 56 and 74 objects, respectively, among 21 000 spectra in 1.75 deg2 [3]. This AGN sample is, by construction, free of morphological or color-selection biases. The measured surface density of BLAGN with IAB ≤24.0 is ∼472±48 deg−2 , a value significantly higher than that of any other optically selected sample of BLAGN with spectroscopic confirmation. Regarding the IAB ≤22.5 sample, ∼42% of them are classified as extended up to z≤1.6, and a large fraction of them are lying close to the color-space occupied by stars in the u∗ -g’ versus g’-r’ color-color diagram, thus leading to up to ∼35% undersestimate of BLAGN when using the classical preselection techniques. The composite spectrum derived from this sample displays a continuum shape which is very similar to that of the SDSS composite at short wavelengths, but it is much redder than that of the SDSS composite at λ ≥ 3000 Å. This observation could be interpreted as a significant contamination from emission by host galaxies (see more details in [3]). The high-reshift galaxy population at 1.5≤z≤5 One of the most striking results of the VVDS concerning the census of highz galaxies is the fact that 970 galaxies have been found with spectroscopic redshifts 1.5≤z≤5. This magnitude-selected sample is 1.6 to 6.2 times larger than previous estimates based on the Lyman-break selection (LBG), with the difference increasing towards brighter magnitudes [11]. The UV continuum in these high-z spectra indicates a vigorous star formation activity of ∼10-100 M yr−1 (Fig. 2). The UV LF of the high-z galaxy population at 3≤z≤4 has been obtained for the VVDS I-band selected sample, and compared to the usual findings based on Lyman-break selections [13] (Fig. 3). The best fit parameters are Φ∗ = 1.24 ± 0.5010−3 mag−1 Mpc−3 and M ∗ = −21.49 ± 0.19, 47 Relative Flux F_lambda The VIMOS VLT Deep Survey (VVDS) Fig. 2. Composite spectra of VVDS I-band selected galaxies in two different highredshift intervals: 2.5≤z≤3.5 and 3.5≤z≤5. assuming a slope α=-1.4. Although Φ∗ is similar to other previous studies, M ∗ is found to be 0.5 magnitudes brighter. As a consequence, the cosmic star formation rate is higher than previously measured at z∼3-4. The unexpectedly large number of bright galaxies found in the I-band selected sample of the VVDS indicates that the usual LBG color-selection technique may be affected by significant incompleteness, even at relatively bright magnitudes. Failures in the color-color identification could be due to imperfect modelling of the stellar emission, AGN contribution, dust absorption and/or intergalactic extinction. Also, the galaxy formation process has to be more efficient that previously assumed at z ≥6-7. 3 Summary and Conclusions This I-band limited sample of the VVDS provides an unprecedented dataset to study galaxy evolution over ∼90% of the life of the universe. New results have already been obtained on the census of high-z galaxies and AGNs, and on the evolution of the global properties of galaxies with cosmic time. The results found on the LF indicate a strong type-dependent and localdensity dependent evolution. The latest spectral types of galaxies are responsible for most of the evolution of the UV-optical LF out to z=1.5. The redshift and luminosity evolution of the colour-density relation supports a galaxyevolution scenario in which star formation/gas depletion processes are reinforced both in the more luminous objects and in high density environments, with a star formation activity progressively shifting with decreasing redshift towards lower luminosity galaxies in low density environments. This behaviour is consistent with a downsizing scenario. The clustering properties measured 48 Pelló et al. w do in on w VVDS z=[3,4] I−band selected sample ti c le G se LB LBG Steidel z=3 Steidel z=4 Sawicki z=3 Sawicki z=3 Fig. 3. From left to right: a) Color-color diagram of VVDS high-redshift galaxies (large dots and triangles), compared to galaxies at other redshifts (small dots). Dashed lines enclose the usual LBG selection region in the color-color diagram. b) UV LF at 3≤z≤4 for the I-band selected sample in the VVDS, compared with the usual values found for LBGs (adapted from Le Fèvre et al. 2005 and Paltani et al. 2006). up to z=1.2, for different spectral types of galaxies, indicate that the most evolved galaxies at this redshift were already located into the strongest peaks. Finally, the VVDS has found a significantly larger galaxy population of galaxies at 1.5≤z≤5 than in previous studies based on classical LBG selection. This result points towards an active star-formation activity at z ≥6-7. References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. Bertin, E., Arnouts, S.: A&A 117, 393 (1996) Cucciati, O., Iovino, A., Marinoni, C. et al: A&A 458, 39 (2006) Gavignaud, I. Bongiorno, A., Paltani, S. et al: A&A 457, 79 (2006) Ilbert, O., Tresse, L., Zucca, E. et al: A&A 439, 863 (2005) Ilbert, O., Lauger, S., Tresse, L. et al: A&A 453, 809 (2006) Le Fèvre, O., Saisse, M., Mancini, D. et al: SPIE 4841, 1670 (2003) Le Fèvre, O., Meiller, Y., McCracken, H.J. et al: A&A 417, 839 (2004) Le Fèvre, O., Ventollani, G., Paltani, S. et al: A&A 428, 1043 (2004) Le Fèvre, O., Ventollani, G., Garilli, B. et al: A&A 439, 845 (2005) Le Fèvre, O., Guzzo, L., Meneux, B. et al: A&A 439, 877 (2005) Le Fèvre, O., Paltani, S., Arnouts, S. et al: Nature 437, 519 (2005) Meneux, B., Le Févre, O., Guzzo, L. et al: A&A 452, 387 (2006) Paltani, S., Le Févre, O., Ilbert, S. et al: submitted to A&A, astro-ph/0608176 (2006) 14. Pollo, A., Guzzo, L., Le Févre, O. et al: A&A 451, 409 (2006) 15. Zucca, E., Ilbert, O., Bardelli, S. et al: A&A 455, 879 (2006) Session II Science with GTC Galaxy Surveys in the Era of Large Ground-Based Observatories R. Guzmán Department of Astronomy, University of Florida, USA, guzman@astro.ufl.edu Summary. I review the observing strategies and recent results of various galaxy surveys over a wide range in redshift using 10-m class telescopes, and compare them to the two major galaxy surveys currently being proposed for the GTC: OTELO and GOYA. Both surveys focus on a region of the observational parameter space that has not been explored extensively yet. I conclude that these GTC galaxy surveys can be indeed highly competitive despite the late arrival of the GTC to the exclusive club of large ground-based observatories, but only if a substantial amount of observing time and resources are allocated for such surveys. 1 Introduction Galaxy surveys at different redshifts provide the necessary data to: • carry out a comprehensive census of the galaxy population in a representative volume of the universe at different epochs. Galaxies provide not only information about the main constituents of the universe but also test particles to trace the properties of the universe as a whole. • study galaxy formation and evolution by providing direct observations to shed light on the formation epoch of the different galaxy components (e.g., measuring the morphology, size, light profile, or bulge-to-disk ratio of galaxies at very high redshifts), the role of the environment (e.g., field, pairs, groups or clusters), the rate and mechanism of assembly of the massive galaxies we see today (via hierarchical mergers or monolithic collapse), the origin of scaling laws (such as, e.g., Tully-Fisher or the Fundamental Plane), the history of star formation and chemical enrichment, or the relation to other astronomical phenomena such as AGN, Gammaray bursts, etc. • probe the large scale structure and the nature of dark matter and dark energy using power spectrum analysis of the three-dimensional distribution of galaxies in the universe, and gravitational lensing by galaxy clusters and galaxy peculiar velocities to map the distribution of dark matter. 52 R. Guzmán • measure cosmological parameters using galaxy properties such as age, size, luminosity, surface brightness, merger rate, or cluster density, as probes to test various cosmological models. One of the main difficulties in galaxy surveys is the problem of getting enough information to assemble a global, representative picture of the universe. Gathering a body of data large and accurate enough to be useful in addressing the broad range of astronomical goals summarized above is the starting point of any major survey. Before the advent of 10m-class telescopes, photometric surveys focused on the study of morphologies, colors, and luminosities gathered samples of as much as ∼ 104 galaxies at z < 0.1, and ∼ 103 galaxies at 0.1z < 1.5. Spectroscopic surveys in turn used samples as large as ∼ 103 galaxies at z < 0.1, and only ∼ 102 galaxies at 0.1 < z < 1.5 to study galaxy ages, star formation rates (SFRs), metallicities, and kinematics. About a decade ago, there were almost no galaxies known at redshifts z > 1.5. With the new generation of large ground-based telescopes and widefield instrumentation, the typical samples of current galaxy surveys, including both photometric and spectroscopic properties, are three orders of magnitude larger than previous surveys over the same redshift range. In addition, the universe at z > 1.5 is now being systematically surveyed with samples of the order of 103 − 104 galaxies including measurements of morphologies, colors, luminosities, ages, SFRs, metallicities, and kinematics. Clearly, we live in an era where observations have a leading edge over theoretical models, and their input is now indispensable to advance theoretical efforts. 2 Overview of current galaxy surveys Over the last decade, several large collaborations have been put together world-wide to carry out the new generation of galaxy surveys. Three main characteristics distinguish this new era of galaxy surveys: (i) guaranteed access over a very long period of time to state-of-the-art wide-field instruments specifically designed for these surveys; (ii) galaxy samples that are 2-3 orders of magnitude larger than any previous work on the field, and can be considered for the first time to be statistically representative of the universe at a given epoch, and (iii) large, homogeneous data sets including a wide variety of photometric and spectroscopic parameters measured to an unprecedented degree of accuracy and over an unprecedented range in redshift. Representative examples of such galaxy surveys in the low redshift regime are the Sloan Digital Sky Survey (SDSS, http://www.sdss.org), which has full-time use of a 2.5-m telescope equipped with a wide-field optical camera and spectrograph, or the Two-Degree Field Galaxy Redshift Survey (2dfGRS, http://www.aao.gov.au/2df/), which has been a key-project on the 4m AAT during 7 years using the 2df optical multifiber spectrograph. Both of these surveys aim at characterizing the properties of the general galaxy population and their relationship with the galaxy environment, and mapping the Galaxy Surveys 53 three-dimensional distribution of matter to a distance of about z = 0.2 using spectroscopic observations of ∼ 106 galaxies down to 19 mag ([9]; [6]). At intermediate redshifts (z < 1.5), there are various major surveys that are currently being conducted. Among them, the Deep Evolutionary Extragalactic Probe (DEEP, http://deep.ucolick.org), and VIMOS VLT Deep Survey (VVDS, http://www.astrsp-mrs.fr/virmos). The DEEP collaboration has had at least 30 nights in 4-m class telescopes to obtain the deep BRIK photometry necessary for the sample selection. In addition, DEEP has been awarded 120 nights at Keck using DEIMOS. The VVDS collaboration in turn has had 45 nights in 4-m class telescopes to prepare for the sample selection, and has been awarded at least 70 nights at the VLT using VIMOS. DEIMOS and VIMOS represent the state-of-the-art in wide-field optical multiobject spectrographs available today in 10-m class telescopes, and were built specifically by the DEEP and VVDS teams, respectively, to carry out these surveys. The main scientific goals of the DEEP and VVDS surveys are very similar: to map the distribution of galaxies, AGN and large-scale structure, and parametrize their evolution over the last 8 Gyrs using spectroscopic measurements of ∼ 105 galaxies down to 24 AB-mag ([12], [14]). A distinguishing aspect of DEEP, however, is their ability to study the internal kinematics of the distant galaxy population due to the higher spectral resolution of DEIMOS. Such internal kinematics provide a powerful new dimension related to the dynamical masses of galaxies. These are intimately tied to dark matter halo masses, which in turn are the fundamental components of galaxies best understood from theoretical simulations. In addition, internal kinematics can be used as a standard volume tracer for testing the various cosmological models and to investigate the origin and evolution of the galaxy scaling laws. Finally, regarding the relatively recent field of galaxy surveys at high redshifts (z > 1.5), it is necessary to mention the pioneer work by the collaboration led by [22], which has been awarded about 30 nights in 4-m class telescopes to obtain the deep optical photometry for the sample selection, and between 75 and 100 nights at Keck and VLT for the optical and near-IR spectroscopic follow-up (http://www.astro.caltech.edu/ccs/). The most ambitious survey of the distant universe to date is the Great Observatories Origins Deep Survey (GOODS, http://stsci.edu/ftp/science/goods). This survey has been awarded a Spitzer Legacy Program, which has provided deep photometry at 3.6 to 24 microns with IRAC/MIPS, and a HST Treasury Program, which has provided deep multi-band optical photometry with ACS. It also has deep Chandra X-ray imaging data, and has assured an “extensive commitment” by ESO and NOAO on 4-m and 10-m class telescopes to obtain deep optical and near-IR photometry and spectroscopy [7]. Both programs use the optical broadband “dropout” technique to identify the Lyman-break spectral feature in galaxies at redshifts z ∼ 3. The data collected so far have been used in a number of pioneering investigations on the nature of the Lyman-break galaxy population at high redshift, their large-scale distribution, their contribution 54 R. Guzmán to the star formation history of the universe, and their relationship to the diffuse intergalactic medium. Fig. 1. Redshift evolution of the comoving stellar mass density (reproduced from [8]). Open symbols show results from the literature at 0 < z < 1 (circle, [5]; triangles, [2]; squares, Cohen 2002; filled squares, HDF-N points). The vertical extent of the boxes shows the range of systematic uncertainty introduced by varying the metallicity and the star formation histories of the mass-fitting models used. The solid and dotted curves show the result of integrating the cosmic star formation history, SFR(z), traced by rest-frame UV light, with (solid) and without (dotted) corrections for dust extinction). The dashed curve shows the integrated star formation history from [17], with their 95% confidence range indicated by the shaded region. A detailed account of all the scientific results achieved by these and other surveys is well beyond the scope of this paper. The readers are referred to the web pages referenced above where they can find a complete relation of all papers published by each collaboration as well as an update of the most recent results. For the purpose of this paper, I would simply like to highlight four key areas of research common to all major current galaxy surveys: Mass Assembly: In current models of structure formation, dark matter halos build up in a hierarchical process controlled by the nature of the dark matter, the power spectrum of density fluctuations, and the parameters of the cos- Galaxy Surveys 55 mological model. The assembly of the stellar content of galaxies is governed by more complex physics, including gaseous dissipation, the mechanics of star formation itself, and the feedback of stellar energetic output on the baryonic material of the galaxies. The total, integrated mass in stars is tightly coupled to the history of star formation, as traced by the infrared background, and to the cold gas content of the universe. Reducing the uncertainties on all of these measurements will provide strong constraints on models for galaxy formation. A summary of the current measurements of the stellar mass density at various epochs is shown in Figure 1 ([8]). Fig. 2. Star-formation density traced by a variety of star-forming galaxies at different redshifts (reproduced from [1]). The open diamonds with error bars are the data points at z = 3 − 4 of Lyman-break galaxies with no dust correction included [21]. The filled triangles and circles are SCUBA galaxies. The two sets of open diamonds with no error bars represent two different estimates of the dust obscuration correction for the Lyman-break galaxies. SFR Density of the Universe: By modeling the “emission history” of the universe at ultraviolet, optical, and near-infrared wavelengths from the present epoch to z ∼ 4, it is possible to answer some key questions in galaxy formation and evolution studies. For instance: is there a characteristic epoch of star and metal formation in galaxies? What fraction of the luminous baryons observed 56 R. Guzmán today were already locked into galaxies at early epochs? Are high-z galaxies obscured by dust? Do spheroids form early and rapidly? Is there a universal IMF? [16]. A summary of the current understanding of the history of the star formation activity of the universe is illustrated in Figure 2 ([1]). Fig. 3. Estimates of metallicity over cosmic time (reproduced from [15]. The heavy solid symbols represent [O/H] metallicities (left axis) at three cosmic epochs derived in a self-consistent way from R23 using the data for the NFGS in [11], [15], and [18]. The light open circles are Damped Lyman Alpha absorption systems and are based on [Fe/H] metallicities (right axis) in the column density weighted analysis of [13]. The evenly spaced open squares are the [Fe/H] age-metallicity relation for the Galactic disk from [24]. The lines represent various theoretical models: the global model from [17] (short-dashed line), the collisional starburst model from [19] (dotted line), and three cuts of overdensity from the numerical simulations of [3] (long-dashed lines). Cosmic Chemical Evolution: The metallicity of the universe and of objects in it provides a fundamental metric reflecting the development of structure and complexity in the universe on galactic scales. This metric is all the more important because it is relatively easily observable and “long-lived” in the sense that heavy atomic nuclei, once produced, are not readily destroyed. The metallicity of a galaxy can only increase monotonically with time (unless large-scale infall of primordial gas is invoked), while other parameters such Galaxy Surveys 57 as the luminosity may increase or decrease depending on the instantaneous SFR. Metallicity is thus less sensitive to variations because of transient star formation events in a galaxy’s history and provides a good tracer of the overall evolution of the stellar populations. A summary of the current estimates of the metallicity evolution as a function of redshift is shown is Figure 3 ([15]). Fig. 4. Characterization of the SDSS power spectrum in terms of constraints on the “shape parameter” hΩm and the baryon fraction fb (reproduced from [23]). The best fit to the power spectrum supports the so-called “concordance cosmology”, i.e., a lambda-dominated universe. Large Scale Structure: The cosmological constraining power of threedimensional maps of the universe provided by galaxy redshift surveys has motivated ever more ambitious programs to measure the shape of the realspace matter power spectrum P(k) as a function of redshift. Analysis of the 2dFGRS and SDSS datasets has corroborated the dark energy-dominated cosmology first suggested by the SN-Ia results and later supported by the WMAP measurements. A summary of all latest measurements of P(k) at various scales is shown in Figure 4 ([23]). 58 R. Guzmán 3 Galaxy Surveys with the GTC When the GTC sees its first-light in 2007, it will have to compete with other 10-m class telescopes that have been fully operational for over a decade. In the previous section, I have emphasised the large amount of work already being undertaken by several international collaborations to survey galaxies over a large redshift range. A fair question to ask is: “can the GTC be competitive in the field of galaxy surveys?”. My answer to this question is a clear “YES!”, but only if the new GTC galaxy surveys are truly unique. In my opinion, this can be achieved if the following two conditions apply: • the GTC galaxy surveys focus on an unexplored region of the observational parameter space. This can be done either by studying different types of galaxies(e.g., selecting the galaxy sample using a completely different selection criteria), or by studying a different set of properties (e.g., observing at a different wavelength range). • the GTC instruments are optimally designed to conduct such new galaxy surveys. There are currently two major galaxy surveys that are being proposed for the GTC which fulfill the two conditions above: OTELO and GOYA. 3.1 OTELO OTELO is a flux limited survey of emission line galaxies down to 8 × 10−18 erg/cm2 /s (S/N=5), in large and well-defined volumes of the Universe using OSIRIS (see contributions by J. Cepa, and J. Gallego in this conference proceedings). The redshift range of interest is 0.24 < z < 6.6, covering a total area of ∼1 square degree. The estimated sample size is ∼ 104 galaxies and AGNs. The main scientific goals of OTELO are: (i) to measure the SFR density of the universe using Hα luminosities at 0.24 < z < 0.5 (cf. Figure 2); (ii) to parametrize the chemical evolution of the universe from z=0.24 to z=1.5 (cf. Figure 3); and (iii) to detect Lyα emitters at z∼5.7 and z∼6.6 [4]). The originality of this survey rests on the sample selection. It uses tunable filters, a key feature of OSIRIS compared to similar instruments in other 10m class telescopes, to do co-moving tomography at a depth that allows to measure fainter emission-line galaxies than those studied by previous surveys, while scanning volumes of the Universe that contain statistical representative samples. OTELO is one of the main science drivers behind OSIRIS, a first generation instrument for the GTC. More information about OTELO can be found in: www.ll.iac.es/project/ osiris/otelo. 3.2 GOYA GOYA (Galaxy Origins and Young Assembly) is an infrared-selected, magnitudelimited survey of the galaxy population at high redshift using EMIR (see Galaxy Surveys 59 contributions by M. Balcells, J. Gallego, and R. Pello in this conference proceedings). The redshift range of interest is z > 1.5, covering an area of ∼0.5 square degrees. The estimated sample size is ∼ 103 galaxies and AGNs. The main scientific goals are two-pronged. Firstly, GOYA aims to provide the first comprehensive study of the nature of galaxies at 2 < z < 3 and assess their evolution over cosmological timescales by comparing directly their rest-frame optical properties with those of the galaxy population in the nearby universe. In particular, the rest-frame wavelength range will provide: (i) direct determinations of the amount of extinction in high-redshift galaxies –one of the most controversial corrections in current studies of the distant universe– using the Balmer decrement technique; (ii) the SFR density if the universe at 2 < z < 3 by measuring Hα luminosities (cf. Figure 2); and (iii) the chemical enrichment of the universe by measuring the Oxygen abundance (cf. Figure 3). Secondly, a most novel aspect of GOYA will be its ability to study the internal kinematics of high redshift galaxies by measuring the emission line velocity widths. As it was mentioned earlier, internal kinematics provide a powerful new dimension related to the dynamical masses of galaxies, which in turn serve as tracers of the dark natter halo masses which are being modelled by the new generation of theoretical simulations [20]. Finally, GOYA will be able to search for primeval galaxies at the earliest epoch of the universe through observations of the OII[3727] emission line up to z=5.4, and Lyα at z > 10 [10]). The originality of this survey rests on the wavelength range of study. At z > 1.5, the rest-frame wavelength range is shifted into the near-IR. Only a near-IR multiobject spectrograph in a 10-m class telescope can provide the efficiency and sensitivity required to carry out a survey like GOYA. To date, such instrument does not exist. GOYA first came to light as the main science driver behind EMIR at the GTC, one of only two near-IR multiobject spectrographs for 10-m class telescopes currently being built in the world. The second such instrument is FLAMINGOS-2 which is being built at the University of Florida for Gemini-S (PI: Stephen Eikenberry). However, the higher spectral resolution of EMIR will allow not only to higher efficiency in observing the emission lines of high redshift galaxies between the forest of OH sky lines, but also to conduct the unique survey of internal kinematics and dynamical masses at a very early period in the history of the mass assembly of galaxies in the universe (cf. Figure 1). Finally, since a large fraction of star-forming galaxies at z > 1.5 behave as “standard candles” following the same scaling law between Hβ luminosity, velocity width, and Oxygen abundance defined by nearby HII galaxies, it is possible to use them to perform the classical redshift-distance test and constrain cosmological models with the maximum discrimination between the various cosmological parameters (Siegel et al. 2004). A more detailed description of GOYA can be found at: www.ucm.es/info/emir/goya. 60 R. Guzmán 4 Conclusions In this paper I have argued that the GTC can still develop a world-class scientific program in the area of galaxy surveys if the proposed GTC surveys focus on an unexplored region of the observational parameter space, and the GTC surveys are optimally designed to conduct such surveys. Two surveys, OTELO and GOYA, are ideally suited to take advantage of the unique capability of the first generation of wide-field instruments at the GTC: OSIRIS and EMIR. However, for any GTC survey to be able to compete successfully with the major surveys that are currently being conducted in all large ground-based observatories, it is essential that GTC grants them a “Key Project” status. All large ground-based and space-observatories today have Key Projects which, in essence, are simply large surveys of faint populations. Survey mode is arguably the most efficient use of high sensitivity and wide field of view characteristic of the new generation of large ground-based telescopes. Indeed, Key Projects are becoming the gold-standard of research in astronomy at the dawn of this XXI century, providing the largest scientific impact and the fastest advance of knowledge in a particular area of research. In my opinion, a successful key project needs: • a scientific program that is both unique and specifically tailored to the characteristics of the telescope and its instrumentation. • guaranteed observing time, essential both to conduct the preparatory groundwork and to carry out the actual project. If the access to the required instrumentation till the completion of the survey is not guaranteed, delays due to technical problems, weather, and changing Time Allocation Committees will jeopardize the survey timely competitiveness and condemn it to failure. • a fast, reliable data reduction pipeline to promptly reduce and analyze the sometimes overwhelming amount of data produced by the new generation of wide-field instruments in survey mode. • adequate resources in manpower, equipment, and funding. • to provide an easy-access, fully-reduced database for use of the entire community. In summary, large ground-based telescopes have been fully operational for over a decade, and several major galaxy surveys are currently underway. In order to compete in this field, GTC will have to make a decisive impact in those areas of research that have not yet been fully explored by other 10m class telescopes. This can be best done by conducting Key Projects that best take advantage of the unique instrumentation of the GTC, such as the OTELO and GOYA surveys. I am convinced that only the scientific return of successful Key Projects will allow GTC to claim its own place among the world-class observatories. Acknowledgements: I am grateful to the organizing committee for the kind invitation and financial support to attend this conference. Galaxy Surveys 61 References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. Barger, A. J., Cowie, L. L., & Richards, E. A.: AJ, 119, 2092 (2000) Brinchmann, J., & Ellis, R. S.: ApJl, 536, L77 (2000) Cen, R., & Ostriker, J. P.: ApJl, 519, L109 (1999) Cepa, J., Alfaro, E., E., Balnd-Hawthorn, J. et al: RevMexAA Ser. Conf., 16, 64 (2003) J. P.: ApJl, 519, L109 (1999) Cole, S., et al: MNRAS, 326, 255 (2001) Colless, M.: Royal Society of London Philosophical Transactions Series A, 357, 105 (1999) Dickinson, M., Giavalisco, M. et al: in The Mass of Galaxies at Low and High Redshift, 324 Dickinson, M., Papovich, C., Ferguson, H. C., & Budavári, T.: ApJ, 587, 25 (2003) Gunn, J. E., & Knapp, G. R.: Sky Surveys. Protostars to Protogalaxies. In: ASP Conf. Ser., vol 43, pp 267 (1993) Guzman, R.: RMxAC, 16, 209 (2003) Jansen, R. A., Fabricant, D., Franx, M., & Caldwell, N.: ApJs, 126, 331 (2000) Koo, D.C.: Bulges, Disks, and Kinematics of Galaxies at z=1. In: Astrophys and Space Science Library, vol 319, ed by D. L. Block et al (Dordrecht: Kluwer Academic Publishers 2004) pp 441 Kulkarni, V. P., & Fall, S. M.: ApJ, 580, 732 (2002) Le Fèvre, O., Vettolani, G., Maccagni, D., Picat, J. P.x et al: ArXiv Astrophysics e-prints, arXiv:astro-ph/0402203 (2004) Lilly, S. J., Carollo, C. M., & Stockton, A. N.: ApJ, 597, 730 (2003) Madau, P., Pozzetti, L., & Dickinson, M.: ApJ, 498, 106 (1998) Pei, Y. C., Fall, S. M., & Hauser, M. G.: ApJ, 522, 604 (1999) Pettini, M., Shapley, A. E., Steidel: ApJ, 554, 981 (2001) Somerville, R. S., & Primack, J. R.: MNRAS, 310, 1087 (1999) Springel, V., White, S.D.M., Jenkins, A. et al.: Nature, 435, 629 (2005) Steidel, C. C., Adelberger, K. L., Giavalisco, M. et al: ApJ 519, 1 (1999) Steidel, C. C., Adelberger, K. L., Shapley, A. E. et al: ApJ, 592, 728 (2003) Tegmark, M., et al: ApJ, 606, 702 (2004) Twarog, B. A.: ApJ, 242, 242 (1980) The GTC 10m telescope: Getting ready for First Light J.M. Rodrı́guez Espinosa, GTC Project Instituto de Astrofı́sica de Canarias, Spain, jmr.espinosa@iac.es Summary. The Gran Telescopio Canarias (GTC) is undertaking a period of intense testing of the control software. Likewise we are undertaking the preparation for the installation of the first optical elements, thereby paving the way towards what we call Technical First Light. This will be done with a small number of segments that will simplify the understanding of the complicated system required for the alignment of the segments. In what follows I will discuss the current status of the telescope, as well as the calendar for the various steps we need to overcome before reaching First Light, and the subsequent scientific operation. Finally, I will mention the status of the science instruments that will be available during the first years of the telescope. 1 Introduction I would like first of all to thank the organisers for the opportunity to present the status of the GTC to the community, and for having dedicated a special session to the GTC and its science instruments. Having said that, in what follows I will describe the current status of the telescope, very briefly I will go through the science instruments, and I will mention the mid term plan to bring the telescope to First Light. Finally, I will also describe the influence on the GTC of the agreement reached with ESO allowing Spain to become a full member of this organization. Other talks will describe in more detail some of the science instruments, so I will spend some time only on FRIDA, the AO fed high spatial resolution instrument, and will say a few words on SIDE, the last instrument being considered, still at a very early stage. 2 Dome Status The telescope enclosure is essentially finished, except for some repair work needed to leave the shutter fully operational. This repair work is a result of the unfinished state in which the contractor left the dome. Our operation 64 J.M. Rodrı́guez Espinosa, GTC Project team has had to change most of the gears (Fig. 1), install chain tensors, and change the shutter motors, to mention just a few of the many parts where important actuations from our team have been necessary. Additionally, the ventilation louvers have had to be sealed to avoid leaks, and the mechanisms for closing these louvers will have to be changed. A solution is however already demonstrated, although its implementation will not be done in the near future. Note that the operation of the louvers is not actually needed for First Light. Fig. 1. Dome shutter chains. The gears in green are new and design to be adjustable to allow the chains to slide sideways dependent on the demands from the shutter 3 Telescope Status The telescope structure is also essentially finished. Work now inside the dome is mostly cleaning, wiring, installing pipes, and repairing some steel surfaces that may have been damaged after the heavy duty work that has happened inside the dome during the past years. Our Control group has started testing the main axes drives (Fig. 2). The last contractor to come to the telescope has been Tekniker, for installing the Instrument Rotators (Fig 3). The two Nasmyth rotators have been installed, and in a month or so Tekniker will visit again to work on the tests of the rotators. Once the control software of the main axes has been tested and the servos properly tuned. This should happen by the end of November, leaving the telescope in a status such that the first pointing tests can be done with a small refractor attached to the elevation axis of the telescope. These tests The GTC 10m telescope: Getting ready for First Light 65 Fig. 2. Overview of the telescope mount. The Control group has started testing the main axes drives Fig. 3. The instrument rotators should allow us to produce a preliminary pointing model of the telescope mechanics without any optics. By the end of November the telescope should be ready to accept the first optical elements. These will be six primary mirror segments, plus the tertiary mirror and finally the secondary mirror. 66 J.M. Rodrı́guez Espinosa, GTC Project 4 Optics Status All optical elements are now at the site. We have received 42 mirror segments (36 + 6 spares), and the final optical quality of the entire mirror is very good, with a root mean squared wavefront error of just 26.5 nm, or a central intensity ratio (CIR) > 90%. The main difficulty we have encountered is the lack of space in the mirror storage room to hold the 42 segments while they are not installed in the telescope (Fig. 4). Meanwhile, we have been performing many tests and dry runs, in order to become familiar with the handling of the mirror segments, and to establish the procedures for installing segments on the primary mirror cell. All this has been done with a dummy segment, loaded with an actual whiffletree, with exactly the same size, and weight as a real mirror segment. Fig. 4. Mirror storage room. The rest of the segments are stored at a temporary storage in another workshop Also, the secondary mirror is now at the observatory (Fig. 5), and its optical quality is excellent, with a CIR > 98%. Its mechanical drives, allowing the slow and fast guiding as well as IR chopping, need to be tuned. This is a difficult operation that will be undertaken in November so that the secondary mirror is ready to be mounted immediately thereafter. The tertiary mirror is also on site and it has been mounted on its cell, and on the telescope for tests. There are also sufficient actuators and edge sensor to start preparing the segments for First Light. The Acquisition and Guiding boxes are also in the mountain, and have been undergoing tests of its many mechanisms, and detector systems. Note that the A&G boxes are sophisticated devices that will allow us to acquire the The GTC 10m telescope: Getting ready for First Light 67 Fig. 5. Secondary mirror. The optical quality is execellent, with a CIR > 98% target, guide the telescope, perform fast guiding in tandem with the secondary mirror, and monitor the quality of the optics in quasi-real time, with the two wavefront sensors and various spatial resolution on the primary mirror available. Fig. 6. One of the three primary mirror segments already aluminised Finally, I’m happy to mention that three primary mirror segments have already been aluminised (Fig. 6). The reflectivity of the coating seems to be very good. 68 J.M. Rodrı́guez Espinosa, GTC Project 5 Science Instruments For First Light, the GTC will have two main instruments, namely OSIRIS and CanariCam, and a backup instrument, ELMER. OSIRIS is a low dispersion multi-object, and tuneable filter imager, being constructed at the IAC under the leadership of Dr. Jordi Cepa. CanariCam is a mid-IR imager and low resolution spectrometer, with polarimetric and coronagraphic capabilities, being built at the University of Florida by its PI, Dr. Charles Telesco. OSIRIS is starting its laboratory assembly, and will start testing the system early next year, once the optical alignment has been finished. CanariCam is currently undergoing laboratory tests, once its assembly and functional characterization has been completed. ELMER is also an imager and low resolution multi-object spectrograph, that has been built at the GTC Project Office, under the leadership of Dr. Marisa Garca Vargas. ELMER has been tested extensively in the laboratory with outstanding results. There are two second generation instruments currently at different stages of completion. EMIR, a wide field near IR cryogenically cooled multi-object spectrograph, is a complicated system being built at the IAC under the leadership of Dr. F. Garzn. FRIDA is the AO fed high spatial resolution integral-field spectrometer, being built in Mexico under the leadership of Dr. A. Lpez. Finally, there is an approved visiting instrument, CIRCE, which is being developed at the University of Florida, by Dr. S. Eikenberry. CIRCE is a near IR camera, with polarimetric capabilities, that is meant to bridge the gap between Day One and the arrival of EMIR. Still in preparation are two additional instruments, NAHUAL and SIDE, which are still in a preliminary stage of development, and plan to undertake feasibility studies before they are presented to the GTC bodies. The PI’s of all of these instruments are presenting talks or posters that will appear in These proceedings, so I will not anything else about them. I will only comment on FRIDA, whose PI is not in this meeting. FRIDA will exploit the Adaptive Optics corrected beam of the telescope. The key element in FRIDA is the integral field unit, which allows, simultaneously, obtaining high spatial resolution and spectral information of fields of over 30”. FRIDA will offer three spectral resolutions, namely R 1500, 4000, and 30000. FRIDA is a collaboration between UNAM (Mexico), the University of Florida (USA), & the IAC (Spain), with an important participation of the UCM. FRIDA is planned to arrive to the telescope by the end of 2010, once the GTC Adaptive Optics system is on line. 6 Mid and Short Term Calendar The current activity at the mountain is mostly software tests of the main axes drives. This will continue for some time, as the tuning of the servos is a delicate and time consuming activity. Some time will have to be given to The GTC 10m telescope: Getting ready for First Light 69 the Instrument Rotator contractor for their tests. These will be done in two periods, during October and November. Once the Rotators are accepted, our control software team will have to install the GTC software in the Instrument Rotators and start a period of tests, this time with the final control software. In December the optics will start being mounted on the telescope. This is also a complex activity, with the additional difficulty of ensuring the security of the mirrors at all times. If all goes as planned First Light would be done before the end of December. After this we will spend a full year commissioning the telescope and the two First Light instruments. This commissioning period consists on making the telescope, axes, optics, A&G, ancillary subsystems, plus control software, to function as a single unit. With the above calendar in mind, Day One, which defines the start of science operation, will commence early in 2008, hence a call fro proposal might be issued late in 2007. 7 Spain membership in ESO and the GTC Spain recently became a full member of ESO. As part of the in kind agreement between Spain and ESO, the GTC will devote 120 nights for science programmes, open to the ESO plus Spain communities. These programmes are to be done with the facility instruments, starting from Day One and till 2011. Besides, there will be 55 additional nights for technical programmes of interest to ESO. It is clear that this in kind payment in GTC nights is bound to increase the pressure for scientific time on the GTC. 8 Conclusion The GTC is now approaching its First Light and subsequent start of science operation. There remain still many tests, and control software debugging, but the time for science is coming. Spain will thus, together with its entrance into ESO, join the league of countries with access to Large Telescope time. Moreover, the amount of telescope time that will be available for Spain should put Spanish Astronomy into the frontier of future discovery. OSIRIS: Status and Science J. Cepa1,2 , M. Aguiar1 , E.J. Alfaro3 , J. Bland-Hawthorn4 , H.O. Castañeda1 , F. Cobos5 , S. Correa1, C. Espejo5 , A. Farah5 , A.B. Fragoso-López1, J.V. Gigante1 , F. Garfias5 , J.J. González5 , V. González-Escalera1, J.I. González-Serrano6, B. Hernández1 , A. Herrera1, C. Militello7 , L. Peraza1, R. Pérez1, J.L. Rasilla1 , B. Sánchez5 , M. Sánchez-Portal8 and C. Tejada5 1 2 3 4 5 6 7 8 Instituto de Astrofı́sica de Canarias, E-38200 La Laguna, Tenerife, Spain, jcn@iac.es Departamento de Astrofı́sica, Facultad de Fı́sica, Universidad de La Laguna, E-38071 La Laguna, Tenerife, Spain Instituto de Astrofı́sica de Andalucı́a (CSIC), Camino bajo de Huetor 50, E-18008 Granada, Spain Anglo-Australian Observatory, P.O. Box 296, 167 Vimiera Road, Epping, NSW 2121, Australia Instituto de Astronomı́a, Universidad Nacional Autónoma de México, Apartado Postal 70-264. México D.F., México 04510 Instituto de Fı́sica de Cantabria (CSIC-Universidad de Cantabria), E-39005 Santander, Spain Acústica y Vibraciones, Departamento de Fı́sica Fundamental y Experimental, Facultad de Fı́sica, Universidad de La Laguna, E-38071 La Laguna, Tenerife, Spain European Space Astronomy Center (ESAC), Apartado 50727, E-28080 Madrid, Spain Summary. OSIRIS is the optical Day One instrument for the 10.4m GTC telescope. OSIRIS will cover the 365 to 1000 nm spectral range, featuring an 8.6×8.6 arcmin field of view, and capabilities for direct imaging, both long slit and multiple object spectroscopy, and fast spectrophotometry. The combination of OSIRIS wide field, tunable filters plus charge shuffling array detectors, will constitute the most powerful instrument for studying faint emission-line sources at any redshift. The present contribution gives an overview of the instrument development, currently in its verification phase before commissioning on site. 1 Introduction After an international Announcement of Opportunity, OSIRIS was selected as the optical Day One instrument for the GTC in 1999. From the very beginning, given the limited number of GTC first generation instruments, OSIRIS was conceived and designed as a multiple purpose instrument, with a wide field of 72 J. Cepa et al. view, optimized in the red but UV-sensitive, with imaging and spectroscopy as the main modes. However, OSIRIS has a special feature that distinguishes it with respect to other cameras and spectrographs: it will be the first common user instrument for 8-10m class telescopes that uses tunable filters [1]. In fact, OSIRIS is optimized for using this filters. 1.1 Tunable filters A tunable filter (TF) is equivalent to a narrow band filter whose central wavelength can be changed within a wide spectral range. For example, the two OSIRIS tunable filters allow covering the full OSIRIS spectral range: the blue OSIRIS TF can obtain narrow band images from 365 through 670 nm while the red TF covers from 620 through 1000 nm. At each wavelength, there is a wide variety of full width half maximums (FWHM) that can be selected by the observer. For the OSIRIS TFs the FWHM can be tuned between 1.2 to 5 nm, where the upper limit depends on wavelength: is larger in the red part of the TF spectral range. Then, where the blue and red OSIRIS TF overlap, an even larger variety of FWHM can be selected. The time to change the central wavelength and the FWHM is of 1 ms, with a tuning accuracy better than 0.01%. Fig. 1. One of the OSIRIS tunable filters. In contrast with conventional interference filters, the TF represents the solution for efficient narrow band imaging in 8-10m class telescopes. Also, TF are equivalent to thousands of interference filters. Then any absortion or emission line can be observed at any redshift. The large plate scales of 8-10m class telescopes and the small pupils required for using filters and grisms small enough to be easily manufactured and at a reasonable cost, imply quite large incident angles of the collimated beam, usually of several degrees. Since the central wavelength of any filter, either OSIRIS: Status and Science 73 tunable or interference, depends on the incident angle, the central wavelength varies across the field of view. This effect makes the conventional interference (i.e.: non tunable) filters unusable for narrow band imaging in large telescopes. Since the TF can compensate this effect by taking different images changing the wavelength tuned, the TF provides the required solution for efficient narrow band imaging in instruments for 8-10m class telescopes. Also, the different combinations of central wavelengths and FWHMs make the TF equivalent to thousands of interference filters. Then any absortion or emission espectral line can be observed at any redshift withouth purchasing a large amount of filters. This flexibility makes OSIRIS a unique instrument that will open a wide variety of scientific cases that cannot be efficently tackled with any other instrument. 1.2 Main OSIRIS characteristics The OSIRIS observing modes are: • • • • • Broad band imaging Narrow band imaging using TF Long slit spectroscopy Multiple object spectroscopy Fast photometry and spectroscopy The main OSIRIS features are summarized in Table 1. Table 1. Main OSIRIS features. The Nod & Shuffle and µShuffle, combined or not with λ-sorting, allow a many-fold increase in the number of targets per exposure. OSIRIS feature Value FOV (imaging) Broad band filters Tunable Filters 8.6×8.6 arcminutes ugriz Central wavelength tunable from 365 to 1.000 nm FWHM tunable between 1.2 to 5 nm depending on wavelength Spectral resolutions 250, 500, 1.000, 2.000, 2.500 and 5.000 for a slit width of 0.6 arcseconds MOS Mask loader capacity of up to 13 user-customized masks About 40 targets per masks using slits or up to several hundred using multiplexing modes1 2 Instrument status The OSIRIS mechanical assembly and the optical alignment of the instrument are already finished. The following step is the verification phase, prior to 74 J. Cepa et al. shipping the instrument to the observatory for the final commissioning at the GTC. In this section an overview of the main characteristics of every subsystem and unit will be given, following the light path from the telescope focal plane to the OSIRIS detector. 2.1 Mask loader The mask magazine has a capacity of 13 masks, either long slit or multipleobject spectroscopy (MOS). The MOS is then achieved by means of usercustomized masks drilled for specific scientific programs. The observer can select the position, size, position angle and shape for every slit. The position of the slits can be derived from OSIRIS images or by astrometric coordinates provided by the user. The field for MOS is of 8.6×5.2 arcminutes. The mask loader allows changing masks in less than 27 seconds in the worst situation. Then this mechanism is much faster than those of other instruments for large telescopes. This allows minimizing instrument overheads. Also, the mask loader guarantees mask position repeatability to minimize target acquisition time. Fig. 2. The OSIRIS mask magazine and mask loader. These mechanisms allow changing masks in less than 27 seconds, which makes this system the fastest within similar instruments for 8-10m class telescopes. 2.2 Filters and grisms wheels OSIRIS has four wheels. Three for conventional filters, with a capacity of 8 filters each, making a total of 24 conventional filters simultaneously loaded, and a fourth wheel with a capacity of 6 grisms and the two OSIRIS tunable filters. The conventional filters avaiable include broad band and TF and spectroscopy order sorters. OSIRIS: Status and Science 75 Grisms and TF can be changed in less than 13 seconds while conventional filters can be changed in 3 seconds, minimizing instrument overheads. All wheels can be moved independently and simultaneously. Fig. 3. The four OSIRIS wheels have a capacity of up to 24 conventional filters for broad band imaging or for sorting TF or spectroscopy orders, plus 6 grisms and the two OSIRIS TF. All wheels move independently and simultaneously. Grisms and TF can be changed in less than 13 seconds while conventional filters can be changed in 3 seconds, thus minimizing instrument overheads. 2.3 Overheads With this mask and filter changing times, OSIRIS will be the most efficient instrument in terms of minimizing overheads. While reading out the detectors, it will be possible to change the mask and the grisms or filters. Then the only overheads will come from the detector readout time, telescope sleewing and target acquisition. As shown in Table 2, OSIRIS has changing times an order of magnitude smaller than those of similar instruments for 8-10m class telescopes. 2.4 Grisms There is a variety of spectral resolutions available for tackling a wide variety of science programs. Resolutions 250, 500, 1.000 and 2.500 cover the full OSIRIS wavelength range. There is a 2.000 grism covering from 400 to 550 nm, specifically designed for stellar population studies at low redshift. The 5.000 grisms cover spectral ranges devoted to specific projects. The highest resolutions, 2.000, 2.500 and 5.000 are based on Volume Phased Holographic grisms (VPH) that provide high diffraction efficiencies (higher than 80%). All resolutions correspond to a slit width of 0.6 arcseconds. 76 J. Cepa et al. Table 2. Measured mean times in seconds for changing the instrument configuration. The figures of the other instruments have been retrieved from the corresponding WWW. Telescope Instrument Mask Grism Filter GTC VLT GEMINI SUBARU OSIRIS VIMOS GMOS FOCAS 20 210 120 120 6 90 90 90 3 80 20 60 Fig. 4. One of the OSIRIS grisms. 2.5 Collimator and folder OSIRIS optics consist in one reflective collimator and a camera. An additional flat mirror is included to fold the beam, resulting in a more compact instrument for their use at the Nasmyth or Cassegrain focal stations to ease scheduling and telescope operation. Both collimator and folder are coated with silver protected coatings of the highest performance. Their reflectivity in the red is larger than 98% while in the UV is larger than 92% at 365 nm. 2.6 Camera The camera is composed of 9 lenses, three singlets and three doublets. It is optimized in the red spectral range but with high UV efficiency. Their excellent image quality, better than specified, allows a full exploitation of the GTC image quality and the best possible seeing of the Observatorio del Roque de Los Muchachos. OSIRIS: Status and Science 77 Fig. 5. OSIRIS collimator during the acceptance procedure prior to its assembly. Fig. 6. OSIRIS camera enclosed in its transport box before their assembly. 2.7 OSIRIS efficieny The combination of the high reflectivity of collimator and folder and the high camera transmission, results in a very efficient instrument, far better than similar instruments for large telescopes, as shown in Figure 7. OSIRIS is more efficient in the red and specially in the optical UV. This spectral compensated efficiency makes OSIRIS a truly multiple purpose instrument. 2.8 Cryostat and detector The OSIRIS detector system is a mosaic of two 2k×4k Matra-Marconi CCDs obtained from the same silicon wafer. Their quantum efficiency (QE) peak at 700 nm with a value of 88%. The QE at 900 nm and 365 nm are 60% and 30%, respectively, thus matching the instrument transmission, red optimized but UV sensitive. 78 J. Cepa et al. 0.9 Instrument Transmission 0.8 0.7 0.6 0.5 0.4 0.3 OSIRIS (GTC) FORS (VLT) GMOS (GEMINI) ELMER (GTC) 350 400 500 600 700 800 900 Wavelength (nm) Fig. 7. OSIRIS efficiency compared with that of similar instruments for large telescopes. OSIRIS is more efficient in the red and in the optical UV. Telescope, detector or filters and grisms are not included. The data for the other instruments have been obtained from the corresponding WWW. The controller allows shuffling charge on the detectors for an excellent sky subtraction in imaging and spectroscopy, using the techniques of differential imaging and Nod & Shuffle spectroscopy. Also, fast photometry and spectroscopy are possible since the frame transfer modes are implemented as well. Fig. 8. OSIRIS detector mosaic in the cryostat. OSIRIS: Status and Science 79 3 OSIRIS user resources Together with the guide for observers, several exposure time calculators will be provided via the WWW: for direct imaging, for tunable imaging and for spectroscopy. Also, a tunable filter performance calculator will allow observers to simulate the desired tunable filter configuration: the FWHM available for the central wavelength selected, and the relative position of sky emission lines. Finally, the mask designer tool will allow the users to select the position, width, length and position angle of every MOS slit, to obtain the file to feed the mask cutting machine. The goal is to be able to prepare the MOS masks during day time from the OSIRIS images obtained the night before. Fig. 9. OSIRIS fully assembled with all subsystems and units in position. 4 OSIRIS Science Program The OSIRIS Science Team is composed of 50 researchers from different countries. The OSIRIS guaranteed time will be distributed among them according to the scientific merit of observing proposals submitted by the Science team members to the Instrument Definition Team (IDT). These proposals will be evaluated by the IDT assessed by external referees. The OSIRIS key scientific project is OTELO (OSIRIS Tunable Emission Line Object) survey. OTELO scientific team is led by OSIRIS P.I. and composed by OSIRIS Instrument Definition Team members plus some invited researchers. The OTELO Project is described in more detail in another contribution to this proceedings. 80 J. Cepa et al. 5 Summary OSIRIS is a multiple purpose instrument of wide field of view, excellent image quality, high red and optical-UV transmission, whose overheads are limited by detector readout and telescope sleewing. These characteristics, together with the use of tunable filters in a common user instrument, and a maximum spectral resolution of 5.000, make OSIRIS the most efficient instrument both in terms of the fraction of photons collected per unit time and the effective observing time per observing nigth, for tackling a wide variety of scientific projects with advantages with respect to similar instruments in 8-10m class telescopes. After the optical alignment, OSIRIS is currently undergoing the verification tests prior to its shipping to the observatory for the commissioning at the GTC Nasmyth focus, and GTC Day One operation. References 1. Bland-Hawthorn, J., Jones, D.H.: PASA 15, 44 (1998) EMIR, the GTC NIR multi-object imager-spectrograph F. Garzón1,2 , D. Abreu1 , S. Barrera1, S. Becerril1, L.M. Cairós1 , J.J. Dı́az1 , A.B. Fragoso-López1, F. Gago1 , R. Grange3 , C. González1 , P. López1 , J. Patrón1, J. Pérez1, J.L. Rasilla1 , P. Redondo1 , R. Restrepo1 , P. Saavedra1, V. Sánchez1 , F. Tenegi1 and M. Vallbé1 1 2 3 Instituto de Astrofı́sica de Canarias, La Laguna, Tenerife, Spain, fgl,dabreu, sbarrera,becerril,lcairos,jdg,afragoso,fgago,cgonzal,plopez,jpr, jperez,jlr,predondo,rrestrep, pablosrp,vsr,ftenegi,vallbe@iac.es Departamento de Astrofı́sica, Universidad de La Laguna Laboratoire d’Astrophysique de Marseille-Provence, Marseille, France robert.grange@oamp.fr Summary. EMIR, currently entering into its fabrication and AIV phase, will be one of the first common user instruments for the GTC, the 10 meter telescope under construction by GRANTECAN at the Roque de los Muchachos Observatory (Canary Islands, Spain). EMIR is being built by a Consortium of Spanish and French institutes led by the Instituto de Astrofı́sica de Canarias (IAC). EMIR is designed to realize one of the central goals of 10m class telescopes, allowing observers to obtain spectra for large numbers of faint sources in an time-efficient manner. EMIR is primarily designed to be operated as a MOS in the K band, but offers a wide range of observing modes, including imaging and spectroscopy, both long slit and multi– object, in the wavelength range 0.9 to 2.5µm. It is equipped with two innovative subsystems: a robotic reconfigurable multi–slits mask and dispersive elements formed by the combination of high quality diffraction grating and conventional prisms, both at the heart of the instrument. The present status of development, expected performances, schedule and plans for scientific exploitation are described and discussed. The development and fabrication of EMIR is funded by GRANTECAN and the Plan Nacional de Astronomı́a y Astrofı́sica (National Plan for Astronomy and Astrophysics, Spain). 1 Instrument description The new generation of 10m class optical and near-infrared telescopes currently under construction, by sounding ever deeper into the Universe, hold the promise of providing, for the first time, a direct view of the processes that shaped the formation stars, galaxies and the Universe itself. Also, they will provide, again for the first time, the capability of detecting and isolating 82 Garzón et al. extragalactic stars and star forming regions with unprecedented sensitivity and resolving power, both spatial and spectral. A collective instrumentation effort is underway to allow these new infrastructures to be used to their full potential. The scientific capabilities of the new telescopes are thought to be enormous, not only because of the larger photon-collecting area, but especially because of the new instruments, which, due to major technological advances, are expected to be orders of magnitude more efficient than their current-day counterparts. In addition, these technological challenges will establish the first steps towards the construction of instrumentation for the forthcoming 30 m+ class telescopes, now at the beginning of their conceptual design phases. The Observatorio Roque de los Muchachos, operated by the Instituto de Astrofı́sica de Canarias (IAC) on the island of La Palma, is the site of the 10 meter Gran Telescopio Canarias (GTC) due for first light in 2007. GTC will be the largest aperture single dish telescope in world. Along this effort, a partnership of Spanish and French research institutions is working on the design and construction of EMIR, an advanced NIR multi-object spectrograph for GTC, which will be visited in this paper. EMIR (Espectrógrafo Multi-objeto InfraRojo, [10, 11]), is a common-user, wide-field camera-spectrograph operating in the near-infrared (NIR) wavelengths 0.9–2.5µm, using cryogenic multi-slit masks as field selectors. Figures 1 and 2 provide the best up to date estimate of the expected performances of EMIR in both observing modes, per spectral band. The most relevant instrumental parameter can be found in Table 1. EMIR will provide GTC with imaging, long-slit and multi-object spectroscopic capabilities. The EMIR Consortium is formed by the IAC, Universidad Complutense de Madrid (UCM, Spain), the Laboratoire d’Astrophysique des Midi Pyrénées (LAOMP, France) and the Laboratoire d’Astrophysique de Marseille-Provence (OAMP, France). EMIR is now at the beginning of its Fabrication and AIV Phase phase, and is due for first commissioning at the GTC in early 2008. This phase is being funded by GRANTECAN and the Plan Nacional de Astronomı́a y Astrofı́sica. EMIR will provide the GTC user community with key new observing capabilities. It is expected that it will be one of the first fully cryogenic multi-object spectrograph (MOS) on a 10m class telescope, hence able to observe in the K band at 2.2µm without the drawback of the high instrumental background common to other conceptually similar instruments. Similar NIR MOS existing or planned for other telescopes are not cooled and reach out to 1.8µm only. Extending MOS capabilities to 2.2µm is the natural next step in MOS design. EMIR will open, for the first time, the study of the nature of galaxies at redshifts beyond z=2 with unprecedented depth and field of view. At these redshifts, the well-studied visible rest-frame of galaxies, in particular the strong H line, is shifted to the K band, allowing key diagnostics of the star formation history of the Universe. EMIR will allow to bridge between the extensive studies at lower redshifts carried out in the nineties on 4m class telescopes and those above z=6 planned for the near future using the far infrared and millimetre wavelengths. EMIR will also provide a link between EMIR, the NIR MOS for GTC 83 Fig. 1. Calculated sensitivities of EMIR in spectroscopic mode, using the actual figures for the optics transmission and the detector quantum efficiency. Dashed horizontal lines indicate 1 and 2 hours of integration time current spectroscopic capabilities and those that will become available when the James Webb Space Telescope (JWST) becomes operational late in this decade. The EMIR design was largely determined by the requirements of its main scientific driver, the study of distant, faint galaxies, the GOYA project [12]. Being a common-user instrument, however, it has been designed to meet many of the broader astronomical community. It is therefore a versatile instrument that will accomplish a wide variety of scientific projects ranging from extragalactic and stellar bodies to interstellar medium and Solar system astronomy. The construction of EMIR pushes the challenges of large-telescope instrumentation to new limits. The GTC 10m aperture translates into a physically large focal surface. Matching the images given by the telescope to the small size of current detectors requires large optics with fast cameras. Large, heavy optics need advanced mechanical design and modelling to bring flexure down to acceptable levels. To work in the region beyond 1.8µm, the EMIR optical system and mechanical structure will be cooled down to cryogenic tempera- 84 Garzón et al. Table 1. Top level specifications of EMIR Wavelength range 0.9–2.5µm Optimization 1.0–2.5µm Observing modes Multi-object spectroscopy Wide-field Imaging Top priority mode K band Multi-object spectroscopy Spectral resolution 5000,4250,4000 (JHK) for 0.006 (3 pixel) wide aperture Spectral coverage One observing window (Z, J, H or K) per single exposure Array format 2048x2048 HgCdTe (Rockwell–Hawaii2) Scale at detector 0.002/pixel OH suppression In software Image quality θ80 < 0.003 Multi-object spectroscopic mode Slit area 60 ×40 , with ∼ 50 slitlets of equal length and width varying between 0.004 and 100 Sensitivity K < 20.1, t=2hrs, S/N=5 per FWHM (continuum) F > 1.4 × 10−18 erg −1 s−1 cm−1 Å−1 , t=4hr, S/N=5 per FWHM (line) Image mode FOV 60 ×60 Sensitivity K < 22.8, t=1hr, S/N=5, in 0.006 aperture tures. Temperature stability and cycle-time requirements pose stringent demands on the design and performance of the instrument’s cryogenic system. A key module of EMIR is a cryogenic mask unit to allow several different configurations of multi-sit masks being available every night, suitable for GTC’s intended queue observing, without warming up the spectrograph. All the afore mentioned aspects need development effort, as the technology is not available or it is not scalable from existing solutions. Finally, we are seeking the development of a documented, robust processing pipeline as an integral part of the instrument and are including such software effort in the developments needed for a successful operation of EMIR. In the subsequent sections we will briefly review the different technical aspects of the EMIR design effort, which are described in full in other papers (see references). It is worth to emphasize again that EMIR is a science driven instrumental project, being its top level design requirements taken directly from the main goals of the GOYA project ([12]). But, at the same time, it is conceived as a powerful and flexible common-user instrument which will open new windows to the community to which it serves ([1]). 2 Optical layout The optical concept of EMIR, ([13, 14]), has been studied from many approaches in order to have a good balance between the performance of the EMIR, the NIR MOS for GTC 85 instrument, the technical risks and the global price. The EMIR requirements make the optical concept extremely challenging, and the design approaches have tried to minimize the trade off between requirements and technical solutions. The parameter that drives mostly the design is the size of the required FOV in both imaging and spectroscopic modes. Requirements such the spectral resolution and operation temperature of the instrument and material availability are also important and have special role in the final design. The optical train, all in transmission, is composed, from end to end, by a cryostat window, acting as a field lens and powered for flattening the GTC focal surface, where the Cold Mask Unit is located. Then a multiple spherical lens collimator, combining a single lens and a triplet forms the image of the GTC secondary at the pupil plane, where the dispersive elements and Lyot stops can be inserted and removed from the beam. A six element camera, all of them spherical except the last one, focus the beam onto the detector after crossing the filter wheel situated between the last camera lens and the detector, mounted on a XYZ movement table. All lenses, including the field lens will be AR coated in the two surfaces. The EMIR optical design is specified for the use of grisms as dispersive components. This option appears to be the most feasible approach, with the strong caveat of the unavailability of such grisms in the market. Technical developments to procure large grisms with high refractive index materials are needed, but no only in the EMIR project, and we have already completed such a development during the PD phase, where a demonstration programme was launched to produce a test sample functional in the K band. This was done in a collaborative effort with the OAMP and the grating manufacturer. The complete dispersive element are formed by a combination of two refractive ZnSe prisms plus the transmission grating, which behaves like a grism as far as the light trajectory is concerned. One key aspect in the development of such a pseudo-grism is the technical quality of the gratings grooves, much deeper in the NIR than in the optical. We already ordered, and have received, one grating specified for each of the atmospheric windows JHK, following the successful results of the previous phase. These components will be sandwiched between two standard ZnSe prisms to form the dispersive element, so called pseudogrism, which will be mounted in the Grism Wheel. This task is being performed by OAMP, and includes the acceptance testing of the grating elements and the design and construction of the mounting barrel, one for each of the three gratings. The full set of optical components of EMIR, except the prisms for the pseudo-grisms, are now being fabricated, after the optical design was fitted to the manufacturer test plate. The camera and the collimator triplet will be delivered mounted on their barrels, being the mounting of the two bigger lenses, cryostat window and the first collimator lens, under the responsibility of the EMIR team. 86 Garzón et al. Fig. 2. Calculated sensitivities of EMIR in image mode, using the actual figures for the optics transmission and the detector quantum efficiency. Dashed horizontal lines indicate 1 and 2 hours of integration time 3 Mechanical concept EMIR will be attached to the mechanical rotator of the Nasmyth-B focus. The mechanical layout of the instrument has been derived from the optical design, taking into account the Nasmyth space envelope. Two flats have been added to bend the beam and a cold bench has been optimized to fulfil the image stability error budget. A mechanical concept has been developed for each subsystem, and a final set of specifications has been obtained to feed the detailed design. After finalising a prototyping phase in which the most critical aspects of the mechanical concept have being tested and qualified, EMIR is now entering into the fabrication and AIV phase. The full details of the mechanical design can be found in [5], [3], [2] and [16]. The EMIR mechanical design relays on the development of a fully cryogenic robotic system which can be remotely reconfigured to form the multi-slit pattern in the instrument focal plane, and which is referred to as CSU along EMIR, the NIR MOS for GTC 87 this paper. To this end, a development contract was run with a industrial partner during the conceptual design phase to accommodate the EMIR needs to feasible technologies, either already existing or due for development in the short term. After the finalisation of that contract, and prior to the launch of the call for the procurement of the final component, we have additionally run a demonstration programme with a different industrial partner devoted to identify the most critical aspects in the mechanism, both in the mechanical and electronics areas. A functional prototype has been produced as a result of this contract. At the IAC we will now intensively and extensively test the prototype, at both room temperature and in cryogenic working conditions, before issuing the definitive call for tenders for the final system. The current status of the mechanical development is described in [16]. It is worth to mention that the project has recently reviewed the results of the prototype test in an Advanced Review Meeting which has concluded successfully. Part of the mechanical items are already being fabricated. In the next months the EMIR project will launch calls for proposals to fabricate the cold bench, the vacuum vessel and the EMIR CSU. These contracts will close all the pending components in the EMIR mechanical supply. 4 Control system The EMIR software and control system [6, 7] is being developed by a multiinstitutional group formed by scientist and engineers from IAC, UCM and LAOMP, under the coordination of the IAC. It follows strictly the prescriptions of GRANTECAN for the development of instrument software, in view of the subsequent integration on the global GTC Control System. EMIR Control is based on a distributed architecture where every subsystem has a self contained objective. The instrument core takes control, synchronizes and triggers all the tasks to carry out a sequence of actions which configure an observation. Here are four main aspects in the control system that have been considered as integral part of the instruments from the beginning: • The EMIR Coordinated Operations, which includes the control of the instrument global configuration related with observations and calibrations. These might have to interact with the GTC control system. It is being built in cooperation by IAC and LAOMP. • The EMIR Data Acquisition System ([8]), which drives the different detector read-out modes and controls the flow of data. It is being developed by IAC, based on a SDSU controller. • The EMIR Observing Programme Management Subsystem ([15]) which is the master programme which monitors the EMIR performances and will ensure an adequate use of the EMIR instrument by the regular astronomers. LAOMP is undergoing its design. 88 Garzón et al. • The Data Reduction Pipeline ([9]), which includes specific filters and reduction packages for each observing mode. It is under the responsibility of the UCM. EMIR is equipped with a Rockwell Hawaii 2 FPA which will driven by a controller based on a SDSU architecture. A second science grade FPA have being tested and accepted at the IAC, using our testing equipment (cryostat plus detector controller) specifically designed and built to this end by the EMIR team at the IAC. The controller is a home made design around the SDSU. We have already completed the first test campaign, [4], in that array, which are summarised as follows: • Gain: 3.03 ± 0.12e−−/ADU • Readout noise: 6.5e− • Well depth: 126, 000 ± 500e−, up to 2% deviation from linearity. • Dark current 0.03/0.15 e− /s @ 77 K. • Maximum pixel rate per channel 140 kHz. All the tests have been performed using a basic version of the final Data Acquisition System to be used in EMIR. The design adopts the final hardware components and architecture and the main software components. The inclusion of auxiliary software capabilities and the system integration is planned for a next phase. To speed up software development a prototype to mimic an observing mode is been developed. This prototype is based as much as possible in standard hardware and software components already available. Coordination of actions required to perform the observing mode, data filters, detector control and data acquisition are aspects covered by this development. 5 Schedule EMIR is now running its fabrication and AIV phase, on which all the EMIR components are in fabrication, or already available to the EMIR team; following this verification and integration and at component, subsystem and system level will result in the final instrument ready to be mounted at the IAC premises ready to be qualified prior to shipping to the GTC. The work is proceeding as expected, with some delays which have been accumulated since the beginning of the project, being the major challenges the procurement of the pseudogrisms needed for the light dispersion and the multi-slit mask subsystem, as described above. With the current development contracts being well underway, or close to be assigned, we are not expecting major impacts on the instrument schedule to completion. Most of the present day uncertainties in the calendar will be fixed before or around the summer 2006, after the signatures of the pending procurement contracts. With all the above in mind, we are now facing an schedule to completion which contemplates four major milestones: • The start of the AIV at component and subsystem level by mid/late 2006. EMIR, the NIR MOS for GTC 89 • The start of the AIV at system level by mid 2007. • The beginning of the commissioning at the GTC by early 2008. • EMIR first light in mid 2008. 6 Scientific exploitation As mention in Sec. 1, EMIR is a science driven project. Two teams are at present working in the early preparation of the scientific exploitation of EMIR ([1]). The GOYA team, aimed at producing at complete census of galaxies in the early Universe, in a epoch of enhanced star formation and the EAST, recently formed, which will cope the non-GOYA topics to conform a coherent Central Program to be developed during the first phases of the instrument at the GTC, by the use of the Guaranteed Time. In addition to the above mentioned efforts, and closely related to them, the EMIR team is going to undertake and intensive and extensive astronomical calibration campaign, cooperating with the GTC team, which will permit to overcome the many problems associated with the more classical ad hoc approach, on which the target measurements are directly compared with data taken on standard stars in more or less similar observing conditions. The spatial astronomical missions have, since long ago, accepted the caveats of such calibration procedures and are developing specific calibration tools which increases the value of the date archives. This calibration problem is particularly important in the queue observing scheme adopted by GTC, which needs the setup of clear and systematic procedures. References 1. Balcells, M.: In Science with the GTC, eds. Rodrı́guez–Espinosa, J. M., Garzón, F. & Melo, V. RevMexAA 16, 69 (2003) 2. Barrera, S., Villegas, A., Fuentes, J. et al: Proc. of the SPIE 5495, 611 (2004) 3. Correa, S., Restrepo, R., Tenegi, F. et al: Proc. of the SPIE 5492, 1331 (2004) 4. Dı́az, J.J., Gago, F., Beigbeder, F. et al: In sdab.conf, 493 (2004) 5. Fuentes, F. J., Sánchez, V., Barrera, S. et al: Proc. of the SPIE 5492, 1319 (2004) 6. López-Ruiz, J. C., Dı́az, J. J., Gago, F. et al: Proc. of the SPIE 4848, 474 (2002) 7. López-Ruiz, J. C., Joven, E., López, P. et al: Proc. of the SPIE 5496, 438 (2004) 8. Gago, F., Diaz, J. J., Redondo, P. et al: Proc. of the SPIE 5492, 1280 (2004) 9. Gallego, J., Cardiel, N., Serrano, A. et al: Proc. of the SPIE 4847, 402 (2002) 10. Garzón, F., Barrera, S. ,Correa, S. et al: Proc. of the SPIE 4841, 1539 (2003) 11. Garzón, F., Abreu, D., Barrera, S. et al: Proc. of the SPIE 5492, 1187 (2004) 12. Guzmán, R.: In Science with the GTC, eds. Rodrı́guez–Espinosa, J.M., Garzón, F. & Melo, V. RevMexAA 16, 209 (2003) 13. Manescau, A., Fragoso-López, A. B., Garzón, F. et al: Proc. of the SPIE 4841, 230 (2003) 90 Garzón et al. 14. Manescau, A., Fragoso-Lopez, A. B., Garzón, F. et al: Proc. of the SPIE 5492, 1735 (2004) 15. Richard, J., Pelló, R., Contini, T. et al: Proc. of the SPIE, 5493, 373 (2004) 16. Sánchez, V., Barrera, S., Becerril, S. et al: Proc. of the SPIE 6269 (2006) CanariCam: Instrument Status and Frontier Science C.M. Telesco Department of Astronomy, University of Florida, Gainesville, Florida, USA, telesco@astro.ufl.edu Summary. CanariCam is the multimode mid-IR camera being developed at the University of Florida for use at the Gran Telescopio CANARIAS (GTC). Here I briefly describe the camera and its key observational modes, and I provide examples of the science that CanariCam will make possible. 1 Overview At the time of this writing, CanariCam has been assembled completely and is undergoing extensive laboratory tests in preparation for formal laboratory Acceptance Tests to be held at the University of Florida in Spring 2007. CanariCam will be shipped to the GTC in early summer 2007, and it will be available for Day-1 science observing. It is functioning well, and we do not anticipate obstacles to our having CanariCam ready for the GTC on Day-1. CanariCam is optimized for use at 8-25 µm, the so-called mid-IR spectral region, but it is useful for certain key engineering observations down to about 2 µm. The goal has been to provide the GTC astronomy community with an outstanding “workhorse” multi-mode instrument for use in the atmospheric windows near 10 µm (extending from about 8 to 14 µm) and 20 µm (extending from 16 to roughly 25 µm). The detector is an arsenic-doped silicon, blocked-impurity-band (BIB, or IBC) device from Raytheon, with peak quantum efficiency (QE) in the 8-25 µm region and a rapid decrease in QE at longer wavelengths. The detector array contains 240x320 pixels, each 0.08”, which provides a field of view on the sky of 19”x26”. The diffraction point spread function is Nyquist-sampled at 8 µm (two pixels per resolution element). The CanariCam science modes available are standard imaging, slit spectroscopy, dual-beam polarimetry, and coronagraphy. All modes are available for use in the 10 µm region, but only imaging and spectroscopy will be available at 20 µm on Day-1. Additional technical information about CanariCam is available in [19]. 92 C.M. Telesco CanariCam will address a broad range of scientific problems. Astronomical bodies at temperatures of 100-1000 K emit significant mid-IR radiation. Of particular importance are the ubiquitous small solid particles-dust-that absorb radiation at virtually any wavelength and transform it into infrared, submillimeter, or millimeter radiation. Mid-infrared continuum emission from the dust is diagnostic of the properties of a great variety of astrophysical objects, including planets, circumstellar disks, star-forming regions, and starburst and active galactic nuclei. With multi-wavelength mid-IR imaging, one can locate energy sources that power often enormous luminosities, trace the distributions of dust particles and their temperatures, and determine how UV and optical radiation, which heats the dust, propagates throughout the infrared-emitting regions. The coronagraphic mode is ideally suited to the investigation of substellar objects in close proximity to “parent” stars. Finally, polarimetric observations allow detailed mapping of the magnetic alignment of dust particles in objects such as circumstellar disks, young stars and active galaxies. Below, I provide additional comments about each mode of operation in the context of examples of the science for which CanariCam will be a valuable tool of exploration. My goal is to provide examples that illustrate CanariCam’s anticipated capabilities, rather than to review the field of mid-IR astronomy. Therefore, for convenience, I have taken the liberty of drawing most of these examples from the research of my colleagues and me, even though much interesting mid-IR research is being carried out by others. In addition, because it is very similar in design to CanariCam (which, however, has additional modes), I present several results from the Florida-built instrument T-ReCS in use at Gemini South, since they are the best examples of what CanariCam will be capable. 2 Imaging Mid-IR imaging is considered to be the fundamental science mode for CanariCam, and implementation of the other modes cannot compromise the imaging performance. At an excellent site like Roque de los Muchachos, the mid-IR point spread function will be dominated by diffraction, not seeing, and the 10 and 20 µm resolutions (λ/D) will often be as good as 0.2” and 0.4”, respectively. Because the design of CanariCam is based on the design of the Florida-built instrument T-ReCS, which is now fully operational at Gemini South in Chile, we have a good idea of how sensitive CanariCam will be-it will be very good [4]! In the broadband 10 µm filter (the N band) on a very good night, we estimate a point-source photometric sensitivity of 0.06 mJy, which is the 1σ noise level achieved in 1 hour of chopped (on plus off source) integration time. On somewhat lower quality nights (i.e., higher background and sky noise, lower transmission), the sensitivity will be worse, but this value gives a good idea of what is possible. Generally, because of much lower atmospheric transmission and correspondingly higher background at longer wavelengths, CanariCam: Instrument Status and Frontier Science 93 Fig. 1. Anticipated CanariCam 10 µm sensitivity for young free-floating brown dwarfs at 20 pc. the 20 µm sensitivity is about ten times worse than at 10 µm. To illustrate the science enabled by this sensitivity, consider the detection limit for “freefloating” brown dwarfs (BDs). Based on model atmospheres [9], we show in Fig. 1 the expected 10 µm flux densities for young (50 and 100 Myr) BDs and giant planets located at 20 pc. The shaded area indicates that young BDs with masses as low as 11-13 M(Jup), where M(Jup) is Jupiter’s mass, can be detected with a S/N > 10 in reasonable integration times. CanariCam will permit exploration of BDs with a broad range of masses [17] and ages, including those with masses near the BD-exoplanet boundary. The outstanding angular resolution of CanariCam can permit fruitful exploration of the detailed properties of many types of astronomical sources. In the area of disk research, CanariCam will be used to search for structure in circumstellar disks where planets are forming or have formed. Planets embedded in both primordial and debris disks influence the disk morphology, and we are thereby able to infer properties of the embedded planetary systems. For example, the well known edge-on disk orbiting the star HR 4796A, which was discovered at 10 µm [10, 12], shows strong asymmetries [18]. A new image at 18 µm, obtained with T-ReCS at Gemini South, is shown in Fig. 2 [7]. The IR emission arises from starlight-heated dust particles. The central clump coincides with the star, and the two outer clumps are ansae associated with the ring. However, the NE lobe is brighter than the SW lobe. This may be related to planetary perturbations [13], or, as suggested for β Pic, it may represent the asymmetric distribution of recently created debris associated with the collisional breakup of a large body [20]. These and other possibilities can be examined with detailed multi-wavelength imaging, a powerful mode of CanariCam. 94 C.M. Telesco Fig. 2. T-ReCS image at 18 µm of the edge-on disk of HR 4796A [7]. Fig. 3. Data profile of disk of the star ζ Lep, at right, with a point source, at left [14]. Even unresolved or barely resolved sources can be of deep interest, and CanariCam on the superb 10-m GTC will provide the best combination of mid-IR resolution and sensitivity of any groundbased observatory. The 230 Myr-old A star ζ Leporis illustrates this well (Fig. 3). This star has a bright disk which was unresolved until we [14] used T- ReCS to show that the disk has a radius of about 3 AU, comparable in size to our solar system’s asteroid belt as had been previously surmised [5]. This appears to be the first-ever CanariCam: Instrument Status and Frontier Science 95 resolution of an asteroid-belt around another star, and this object may be the archetype for a new class of disk, the asteroid-belt analog. Another excellent example of the power of high angular resolution is the imaging of the active galactic nucleus (AGN) of the Circinus galaxy, which Packham et al. [15] showed must be smaller than 0.2”, or 4 pc, at 10 µm, a size that is a challenge to our understanding of AGNs in the context of the unified model. 3 Spectroscopy The CanariCam spectrometer sub-system follows that of the Czerny-Turner layout using any one of four classical plane gratings installed on a turret. The general grating properties are listed in Table 1. The indicated diffractionlimited resolving powers are approximate, with final values to be determined during the laboratory tests. By assuming that the limiting fluxes scale as R1/2 , where R≡ λ/∆λ, we can estimate sensitivities in this mode from the broadband sensitivities. The approximate Lo-Res-10 and Hi-Res-10 continuum point-source sensitivities (1σ, 1 hour, chopped integration) are 1 mJy and 3 mJy, respectively, and the corresponding line sensitivity in the 10 µm region is about 10−18 Wm−2 . Table 1. CanariCam spectroscopy Grating Lo-Res-10 Lo-Res-20 Hi-Res-10 Hi-Res-20 Spectral range (µm) 8-14 16-26 8-14 16-26 R 175 120 1310 890 Among the exciting problems that can be addressed with CanariCam’s spectroscopic mode is the study of the mineralogy and composition of large molecules and solid particles in many types of astrophysical environments. By determining the 10 µm spectrum at many locations very near the Circinus AGN, Roche et al.[16] were able to show (Fig. 4) that features associated with polycyclic aromatic hydrocarbon (PAH) molecules are weaker at the AGN core than in surrounding regions, presumably indicating that the AGN core is embedded in a much more extended region of star formation also manifested by a silicate absorption feature. Silicate minerals exhibit broad silicate features at both 10 and 20 µm, which can be in either absorption or emission. The shapes of these features in many types of astrophysical sources differ from that of the general interstellar medium. For example, the disks of Herbig Ae/Be stars (pre-main-sequence A stars), as a class exhibit a rich variety of shapes of the 10 µm emission feature, some of which are more peaked, like that of 96 C.M. Telesco Fig. 4. Spectra at 10 µm of the silicate feature at various locations near the Circinus AGN [16]. Fig. 5. Silicate emission features in SVS20-N & S, as well as (top) the composite spectrum [6]. amorphous silicate dust in the ISM, some of which are nearly flat-topped, and some of which have sub-structure indicative of crystalline components [21]. This variety may represent evolutionary trends or other fundamental properties of the sources, but their study offers the opportunity to probe the solid particles in complex environments, including those where planets are forming. As an illustration, we show in Fig. 5 silicate spectroscopy of the class 0/1 protostars SVS20 N & S located at 250 pc. These deeply embedded stars, which are only about 0.1 Myr old and separated from each other by about 1”, exhibit very different silicate features. One (SVS20-S) is in emission, and the other is CanariCam: Instrument Status and Frontier Science 97 in absorption [6]. We do not yet know why the features from these two coeval stars appear so different, but clearly this spectral information provides one basis for detailed modeling of the radiation transfer in this complex environment. Interestingly, they also show evidence for the crystalline sub-structure, which therefore can appear very early in the evolution of the circumstellar dust. A variety of diagnostic emission lines are also available in the mid-IR spectral region. These include the H recombination lines at 12.4 µm (7-6) and 11.3 µm (9-7) and key fine-structure lines, most importantly the [ArIII] 9.0 µm, [SIV] 10.5 µm, and [NeII] 12.8 µm lines. While the CanariCam spectral resolutions are not high enough to determine much dynamical information from the line spectra, they can tell us much about the distribution of excitation in interesting astrophysical environments. For example, the indicated finestructure lines are a powerful probe of the UV continuum and therefore the star-forming complexes in starburst galaxies. Due to their inherently high extinction, these regions are often heavily obscured visually and even in the near-IR, so these mid-IR lines are sometimes the only way to accurately assess the magnitude and distribution of central star formation. The classic case is M82, which has the added problem of being edge-on, which greatly increases the line-of-sight extinction. Mid-IR continuum and line imaging is one of the few ways to probe this archetypal galaxy’s starburst core [11, 1]. 4 Polarimetry CanariCam will have the first dual-beam mid-IR polarimetric mode. Initially it will only be used in the 10 µm region, but, at a later date, it will be possible to extend this capability to the 20-micron region with the implementation of the corresponding half-wave plate. Since this is a very unique capability, the design warrants further description. The key components of the polarimetric design are: (1) a cooled, rotatable (sulphur-free) CdSe half-wave plate (HWP, retarder) within the cryostat located just upstream from the telescope focal plane; (2) a focal-plane mask at the telescope focal plane; and (3) a (Sulphurfree) CdSe Wollaston prism (analyzer). The HWP will be rotated sequentially to four different discrete orientations (0, 22.5, 45, and 67.5 degrees), with images being taken at each HWP orientation. The Wollaston prism, which, to our knowledge, is the largest ever built, is inserted into the beam on a slide, and produces an angular separation between the orthogonally polarized states, thereby producing two beams, the so-called o and e rays, which results in two images of the object being formed on the detector. This simultaneous measurement of the ordinary (o) and extraordinary (e) rays not only increases observational efficiency but also minimizes effects of seeing and changes in atmospheric transparency. When using a dual-beam analyzer, a special focal plane mask is required so that extended objects can be observed without overlap of the orthogonally polarized images. The separation of beams is usu- 98 C.M. Telesco ally a compromise between possible optical aberrations produced for large separations and cross-talk for too small a separation. Large separations are convenient, since extended objects may be fully covered by one of the mask gaps, and observations can be made with a single setting of the telescope. For a dual-beam polarimeter, an absolute uncertainty in the degree of polarization of 0.5% requires a S/N ratio of ∼300:1 in total flux. For the source-limited case this corresponds to 8 × 104 photons or 4 × 104 per Stokes parameter. Thus, with a dual-beam system the accuracy obtained is a function of photon numbers only, and accurate polarimetry of bright sources can be carried out during observing conditions that are too poor for almost any other type of quantitative observation. CanariCam’s polarimetric mode will be able to measure degrees of polarization as small as ∼0.1% in both the 10-micron atmospheric window. Table 2 gives an indication of the expected CanariCam polarimetric sensitivities. The table indicates the level of polarization P that can be measured with S/N ≈ 3 for a source with 10 µm flux density Fν . Table 2. CanariCam polarimetric sensitivity Fν (10 µm) P(%) 1σ(%) 10 mJy 50 mJy 165 mJy 1.5 0.3 0.10 0.5 0.1 0.03 CanariCam will be a pioneering instrument in mid-IR polarimetric studies. While not entirely negligible in the mid-IR, polarization due to scattering from dust, which is an important polarization process at shorter wavelengths, is expected in most astrophysical environments to be swamped by transmission through, and emission by, populations of elongated particles. An elongated dust particle both absorbs and emits the electric-field component parallel to its long axis. For a population of elongated grains that are absorbing radiation from a more distant mid-IR emitting source as well as emitting their own midIR radiation, multi-wavelength mid-IR polarimetric measurements can permit one to distinguish the absorbed and the emitted components [3]. As the pioneers of this field such as Dave Aitken and Jim Hough have demonstrated, exciting science is possible with mid-IR polarimetry. Because magnetic fields align elongated grains, the resultant mid-IR polarization distribution that we may determine for young circumstellar disks has the potential to provide tremendous insight into the distribution of the magnetic fields in these environments [2]. These magnetic-field distributions must play a critical role in the formation of planets in these systems. While progress is being made [2] in modeling the expected mid-IR polarization distribution for a specific magnetic-field configuration, much work in this area needs to be done. Hope- CanariCam: Instrument Status and Frontier Science 99 fully, the CanariCam polarization mode will serve as a great stimulus to this important area of research. As an excellent example of the value of mid-IR polarimetry, consider the study by Fujiyoshi et al. [8] of the compact HII region G333.6-0.2. Their multiwavelength polarimetry across the 10 µm region permitted them to separate the emission and absorption components. By considering the polarized emission, they then inferred the distribution of the magnetic field (Fig. 6). They conclude that the magnetic field curvature and strength is well explained by a model in which the magnetic field has been compressed by the wind from a young star, which in turn has led to the compression of adjacent cloud material to produce a density enhancement that may in fact be a future region of star formation. This beautiful piece of work shows the insight that is uniquely provided by polarimetric studies. Fig. 6. Mid-IR polarization of the compact HII region G333.6-0.2 [8]. Left panel: absorption-subtracted 12 µm polarization vectors, rotated by 90◦ . Right panel: the inferred magnetic field structure and its relationship to the expanding wind and adjacent clump compression. 5 Coronography For stars observed in the mid-IR, thermal background emission from the sky and the telescope will be many orders of magnitude larger than the stellar flux. By the use of chopping and nodding techniques, however, it is possible to remove this background at levels approaching one part in a million. Once the background has been removed, the focal plane intensity will be dominated by the stellar point spread function (PSF), the wings of which can be thought of as a “halo” arising from diffracted and scattered light. The key motivation for the coronagraphic mode is to suppress the stellar halo, or PSF wings, to allow circumstellar searches for disks and faint companions. An important goal of an effective coronagraphic mode is to minimize residual diffractive structure in the focal plane with minimum losses in field of view and throughput. 100 C.M. Telesco Fig. 7. Azimuthal averages of the point-spread function with and without the coronagraphic mode. To illustrate key performance advantage of the CanariCam coronagraphic mode, we consider the case of good atmospheric conditions at an observing wavelength of 10 µm defined by a Fried scale length of order the size of the aperture (10 m) and an outer scale twice the size of the aperture (20 m). These parameters define peak atmospheric conditions. We assume that the star is occulted in the telescope focal plane by a hard-edged (top-hat), low reflectivity circular occulting mask 0.83 arcsec in radius. In Fig. 7 we show a plot of the log of the azimuthally averaged intensities as a function of the distance from the star. The top line represents the stellar image PSF for no coronagraphic masks. The lower line represents the stellar image for the occulting mask and a Lyot mask that has a serrated and hex-shaped outer mask that matches the pupil shape, including masks that block the secondary-mirror spiders. We find that one achieves a suppression ratio of about an order of magnitude. The baseline CanariCam coronagraphic mode employs a rotating Lyot stop and maximizes the throughput but provides some leeway for the operational complexity anticipated as the mask rotates. The basic design, then, consists of: (1) a hard-edged (top hat), low-reflectivity, focal-plane (i0) mask 0.83 arcsec in radius; (2) a hard-edged, rotating Lyot stop with a spider mask with widths 20 times the spider-image width. The rotating Lyot stop is dodecagonal (12-sided) and scaled from the input pupil so that the outer dodecagonal edge is 90% the size of the image of the original. A central hard-edged, circular mask blocks out the secondary mirror/obscuration; that mask is 140% the size of the image of the original. The total throughput of the Lyot stop is 66%. CanariCam: Instrument Status and Frontier Science 101 Fig. 8. The horizontal dashed line indicates the detection limit of CanariCam in the coronagraphic mode for you giant exoplanets. Based on experience at shorter wavelengths, the sensitivity of the coronagraphic mode is expected to be dominated by systematic effects. Consider the following. To detect a faint companion around an occulted star one must subtract from the program object’s observed intensity profile that of a comparison star observed in the same coronagraphic configuration. This subtraction of the normalized profile can be made to an m (fractional) level of accuracy (i.e., probable error equals m times the intensity at the radius where the companion may be located). Experience in the near-IR suggests that m is approximately 0.1-0.2. Therefore, if the coronagraphically suppressed profile is about ten times fainter than the standard (non-coronagraphic) imaging profile, the absolute error in the measured flux is about ten times smaller than for standard imaging. In some cases this permits one to be background-limited rather than profile-subtraction-limited. Assume, for example, that m = 0.2. For 1 h of chopped integration for the program object and another 1 h for the profile object, the 5σ detection limit, at a companion located 20 AU (1”) from a Sun-like star 20 pc away, is about 0.4 mJy in coronagraphic mode compared to 1 mJy in the standard imaging mode. As illustrated in Fig. 8, this seemingly modest gain now permits one to potentially detect young giant exoplanets with masses smaller than 10 M(Jup). CanariCam will be a tremendous resource for the GTC community. I hope I have been able to convey at least some sense of the exciting science that will be possible, and I thank you very much for giving me the wonderful opportunity here in beautiful Barcelona to tell you about CanariCam. 102 C.M. Telesco Acknowledgements: It is with pleasure that I acknowledge my fellow team members (past and present) of the CanariCam instrument team who have worked, and continue to work, so hard to bring CanariCam to the GTC: Chris Packham, Jeff Julian, Kevin Hanna, Frank Varosi, Roger Julian, David Hon, Craig Warner, Dave Ciardi, Christ Ftaclas, Jim Hough, Margaret Moerchen, Robert Piña, Jim French, Glenn Sellar, and Mark Kidger. References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. Achtermann, J., Lacy, J.: ApJ, 439, 163 (1995) Aitken, Efstathiou, McCall, & Hough: MNRAS, 329, 647 (2002) Aitken, D., Hough, J., Roche, Smith, C. Wright, C.: MNRAS, 348, 279 (2004) Bouchet, P., DeBuizer, J.M., Suntzeff, N. B., Danziger, J., Hayward, T.L., Telesco, C.M., Packham, C.C.: ApJ, 611, 394 (2004) Chen, C., Jura, M.: ApJ, 560, L171 (2001) Ciardi, D. R., Telesco, C. M., Packham, C., Gómez-Martı́n, C., Radomski, J. T., De Buizer, J. M., & Phillips, C. J.: ApJ, 629, 897 (2005) Fisher, R., Telesco, C., Knights, S., Volk, K., Packham, C.: in prep. (2006) Fujiyoshi, T., Smith, C., Wright, C., Moore, T., Aitken, D., Roche, P.: MNRAS, 327, 233 (2001) Hubbard, W., Burrows, A., & Lunine, J.: ARA&A, 40, 103 (2002) Jayawardhana, R., Fisher, S., Hartmann, L., Telesco, C., Piña, R., Fazio, G.: ApJ, 503, L79 (1998) Jones, B., Rodrı́guez-Espinosa, J. M.: ApJ, 285, 580 (1984) Koerner, D., Ressler, M., Werner, M., Backman, D.: ApJ, 503, L83 (1998) Liou, J., Zook,H.: AJ, 118, 580, (1999) Moerchen, M., Telesco, C., Packham, C., Kehoe, T.: ApJ Letters, in press (2007) Packham, C., Radomski, J. T., Roche, P. F., Aitken, D. K., Perlman, E., AlonsoHerrero, A., Colina, L., Telesco, C.: ApJ, 618, L17 (2005) Roche, P. F., Packham, C., Telesco, C. M., Radomski, J. T., Alonso-Hererro, A., Aitken, D. K., Colina, L., Perlman, E.: MNRAS, 367, 1689 (2006) Sterzik, M., Pascucci, I., Apai, D., van der Bliek, N, Dullemond, C.: A&A, 427, 245 (2004) Telesco, C. M., Fisher, R. S., Piña, R. K., Knacke, R. F., Dermott, S. F., Wyatt, M. C., Grogan, K., Holmes, E. K., Ghez, A. M., Prato, L., Hartmann, L. W., Jayawardhana, R.: ApJ, 530, 329 (2000) Telesco, C., Ciardi, D., French, J. et al: Proc. of SPIE 4841, 913 (2003) Telesco, C. M., Fisher, R. S., Wyatt, M. C., Dermott, S. F., Kehoe, T., Novotny, S., Marias, N., Radomski, J. T., Packham, C., De Buizer, J., Hayward, T. L.: Nature, 433, 133 (2005) van Boekel, R., Waters, L., Dominik, C., Bouwman, J., de Koter, A., Dulllemnd, C., Paresce, F.: A&A, 400, L21 (2003) Session III S.E.A. prizes Radiative Transfer in Molecular Lines. Astrophysical Applications A. Asensio Ramos Instituto de Astrofı́sica de Canarias, 38205, La Laguna, Tenerife, Spain, aasensio@iac.es Summary. This paper presents a short summary of the work carried out during my doctoral Thesis. It presents the development of a variety of methods and techniques for solving radiative transfer problems in molecular lines, and their application to some research problems in molecular astrophysics. 1 Introduction In 1926, Sir Arthur Eddington affirmed that “only atoms are physics, molecules are chemistry”, advising in this way his astronomy colleagues of not wasting their time trying to find molecular species in the Universe. Curiously, only a decade after, several spectral lines were detected in the wavelength range between 3900 and 4300 Å [18, 19] which remained unidentified until several later works demonstrated that they are produced by electronic transitions in CH molecules, which give rise to the so-called G-band. Other lines were assigned to CH+ . This discovery suddenly changed some of our ideas about the Universe, since the existence of molecular species was considered to be restricted to the Earth. After the discovery of the CH lines, molecular astrophysics became one of the most exciting and prolific branches of modern astrophysics. Even more striking was the detection of rotational CN lines resulting from transitions between its lowest energy levels. Their rotational temperatures were very close to 2.3 K and independent of the line-of-sight. CN has a very high dipolar moment and collisions barely affect the excitation state of the rotational levels. On the contrary, the excitation state of the rotational lines are mainly driven by the radiation field illuminating the molecules, being almost in equilibrium with this radiation field. Therefore, this result suggested the presence of an isotropic radiation field at a temperature of ∼2.3 K. This result, considered at that moment of quite limited interest for scientists like the Nobel laureate G. Herzberg, was confirmed twenty-five years later when the Cosmic Background Radiation (CMB) was discovered [29]. It became then 106 A. Asensio Ramos obvious that the results obtained from the CN lines are produced by the radiative excitation due to the remnant black-body radiation after the Big Bang. The optical observations at that time did not lead to any other detection of molecular species in the Universe. The discovery of new molecular species had to await the development of the first radiotelescopes. The new spectral region opened up by these telescopes produced a huge amount of new information. The rotational transitions of many molecules were expected to be situated in the radio spectral range and the presence of many molecular species (e.g., OH, H2 O, NH3 , H2 CO) was confirmed, even by the first low-sensitivity telescopes. During the seventies, and thanks to the expansion of the wavelength range covered by the receivers in the radiotelescopes, other molecules could be discovered. A very strong emission was observed at 2.6 mm towards the Orion nebula which was assigned to the J=1-0 rotational transition of CO [39]. In fact, this molecule is one of the most widely used molecular diagnostics. Other detected species were CS, HCN, CH3 OH and HCO+ . The study of the cold Universe could then be accomplished. The observed molecular lines allowed the exploration of the physical conditions in different astronomical objects by using different molecular species. The symbiosis with molecular spectroscopy turned out to be fundamental. Many of the molecular species were very difficult to produce in the laboratory due to the special conditions of low density and high radiation fields present in the interstellar medium. Therefore, the frequencies of the transitions and the molecular properties were unknown. The astrophysical objects selected for the observations of molecular species were considered like molecular spectroscopy laboratories where new species were found and which had to be identified by their spectral lines. All these steps led to a new branch in Astrophysics known as Molecular Astrochemistry. Its main aim is to answer the question of how molecular species are formed in such low density and highly irradiated environments and to obtain information on the physical properties of the medium. Additional important advances in Molecular Astrophysics were feasible thanks to the launch of the Infrared Space Observatory (ISO) satellite. The two spectrometers onboard ISO, the Short Wavelength Spectrometer (SWS) and the Long Wavelength Spectrometer (LWS) have been used to detect several molecular species and to show that some molecular species are found in many different systems. The detection of CH3 [21], an important precursor in the development of the hydrocarbons chemistry or the striking detection of benzene in the envelope of an evolved star [16] can be considered as milestones of ISO. Future IR and sub-millimeter satellites like Herschel1 and ground-based interferometers like ALMA2 will expand our knowledge of the molecular species present in the Universe. 1 2 http://www.esa.int/science/herschel http://www.eso.org/projects/alma Radiative Transfer in Molecular Lines. Astrophysical Applications 107 Interestingly, molecular species are also present in the solar atmosphere, which is a sufficiently dense and cool medium to allow the efficient formation of molecules. The majority of the molecules found in the solar atmospheric plasma are diatomic, although in the cooler sunspots, even water has been detected [38]. This Thesis focuses on some key problems in the field of molecular astrophysics, with emphasis on developing the radiative transfer tools that will be needed to scientifically exploit the future observations we will be able to obtain with Herschel, ALMA, Gran Telescopio Canarias3 (GTC), GREGOR4 and the Advanced Technology Solar Telescope5 (ATST). In addition, by using the Sun as a unique molecular physics laboratory, this Thesis aimed also at making a significant contribution to the emerging field of molecular spectropolarimetry. 2 Radiative Transfer Tools Molecular lines contain key information on the physical properties of the cool regions of the Universe. For this reason, it is of crucial importance to be able to model the observed spectral line radiation. Molecules are found in the stellar envelopes of evolved stars and in the interstellar medium. Therefore, they are usually immersed in strong radiation fields, coming from the lower parts of the atmospheres or from the UV ionizing photons of young and massive stars. This radiation field excites the molecular levels driving them far from thermodynamic equilibrium. In order to infer correctly the physical properties of the astrophysical plasma under consideration, it is crucial to take into account that the lines are typically formed outside local thermodynamical equilibrium (LTE) conditions. The most efficient numerical methods developed for the solution of radiative transfer problems in stellar physics [33] have recently started to be applied to the case of radiative transfer problems in molecular lines [1]. Previous schemes of solution were based on Montecarlo methods, which suffer from some well-known problems like statistical noise. A recent work oriented towards developing test problems for the new RT codes has shown that the techniques used in stellar astrophysics are quickly been introduced in the field of molecular astrophysics [37]. We show in this Thesis that the fastest numerical methods developed so far can be applied to radiative transfer problems in molecular lines. An example of a research field in which molecular astrophysics could play a fundamental role is in the determination of chemical abundances in metal poor stars. One of the interesting problems which could be investigated with the RT tools presented in this Thesis is the “enigma” of the oxygen over-abundance 3 4 5 http://www.gtc.iac.es http://gregor.kis.uni-freiburg.de/ http://atst.nso.edu 108 A. Asensio Ramos Population correction Next point Population correction Formal solution Formal solution Fig. 1. Schematic difference between the Accelerated Λ-iteration (MALI) method of Rybicky & Hummer [30] and the Gauss-Seidel (MUGA) iterative schemes of Trujillo Bueno & Fabiani Bendicho [33]. The MALI scheme consists on calculating the radiation field for all the points and then the population correction is carried out. On the other hand, in the MUGA scheme, once the radiation field is known at the spatial point under consideration, the population correction is carried out and then we move to the next point in the atmosphere taking into account the previously corrected populations. in metal poor stars [11]. This is crucial for the determination of the age of many astrophysical objects like the globular clusters by means of their oxygen enrichment. The oxygen abundances obtained with several different tracers are not in good agreement [10]. Among these tracers, we have the OH lines in the UV. Two fundamental problems arise, which are intrinsic to the technique used for obtaining the chemical abundances. This technique is based on a comparison between spectroscopic observations and synthetic spectra obtained in different atmospheric models. The chosen atmospheric models, typically onedimensional and in radiative equilibrium, and the approximation employed to obtain the molecular abundances, may seriously influence the emergent spectrum. With the recent development of realistic three-dimensional simulations, the situation for the calculation of atomic abundances has changed and many of the results have to be revised. As an example, the iron abundance in the solar atmosphere has been recently revised using NLTE synthesis in threedimensional hydrodynamical models of solar surface convection including the effect of the radiation transfer in the energy budget equation [32]. For the first time, it is found that NLTE spectral synthesis in the 3D hydrodynamical models of the solar atmosphere yields the meteoritic iron abundance. After such new developments, it is now of possible interest to investigate the impact of non-LTE and chemical non-equilibrium effects when obtaining molecular abundances. We have developed the tools needed for investigating this and other type of problems. However, two important obstacles still persist: the enormous lack of reliable molecular data and the need of computing resources. The first problem has a difficult solution in view of the amount of available molecular Radiative Transfer in Molecular Lines. Astrophysical Applications 109 data and of new observational data. Particularly urgent is the need of stateto-state collisional rates for molecular lines. There are some calculations of collisional rates for the ground levels of some molecules, which are of interest in very cold media. However, collisional rates for excited vibrational states and between different electronic states are not known. Even an approximate estimation of such collisional rates would be of enormous help for accounting for non-LTE effects in molecular lines. The second problem is much more related to technology and efforts in parallelization of non-LTE codes [24]. We describe in detail in the Thesis the computer code that we have developed for the solution of radiative transfer problems in molecular astrophysics assuming spherical geometry [9]. We generalize fast iterative methods based on the Gauss-Seidel and Successive Overrelaxation [33] to spherical geometry with macroscopic velocity fields. We show that the fundamental properties of these iterative methods are maintained when spherical geometry is considered. A schematic representation of the fundamental difference between the old and the new methods is presented in Fig. 1. The convergence rate obtained with these methods is much higher than that of the previous methods with a negligible increase in the computational time per iteration. We show two applications of the computer code. The first one concerns the formation of pure rotation water spectral lines in a hot shell of the molecular complex SgrB2 [17] and the second one concerns the formation of CO vibration-rotation lines in the envelope of the red supergiant VY CMa. 2.1 Chemical Evolution Molecules are usually found in highly dynamic systems (e.g., the solar atmosphere, winds of AGB stars, ...) and their formation is influenced by the time variation of the physical conditions in the medium. In this case, it is not correct to use the assumption of instantaneous chemical equilibrium (ICE) and it is fundamental to consider all the reactions which create and destroy any given species and solve the full chemical evolution problem. We have developed a computer code that solves the stiff differential equations that describe the chemical evolution problem. A very interesting problem we have tackled in this Thesis that relied on the application of our chemical evolution code is the study of the temporal evolution of the carbon monoxide abundance in the solar atmosphere. The objective of this investigation was the resolution of the “enigma” emerged 30 years ago when [28] inferred very low brightness temperature from their discovery of strong ro-vibrational CO lines at 4.7 µm observed close to the edge of the solar disk. It was then suggested by [12] that the low chromosphere might not be hot at all but could instead be permeated by CO-cooled “clouds” at altitudes between 500 and 1000 km above continuum optical depth unity. This led to controversy because other diagnostics had suggested the existence of a uniformly hot chromosphere with a minimum temperature of about 4400 K near 500 km and a temperature rise above this temperature-minimum region. Over the last few years, it has 110 A. Asensio Ramos Fig. 2. Solid line: Height variation of the time-averaged CO concentration obtained from the chemical evolution calculation in the strongly dynamic simulation case. Dashed line: Time-averaged CO concentration corresponding to the ICE approximation, but calculating the CO number densities of the atmospheric models associated to each time step by using the same chemical evolution code until reaching the ensuing equilibrium concentrations. Dotted line: Time-averaged CO concentration corresponding to the ICE approximation, but calculating the CO concentrations directly from the Saha chemical equilibrium equations. A comparison of the dashed and dotted lines illustrates the reliability of the chosen database for the chemical evolution calculations. In any case, in order to be fully consistent with our comparisons, all ICE results refer to “evolution until equilibrium” calculations become increasingly evident that for understanding the thermal structure of the solar chromosphere we need a rigorous investigation of the CO formation and destruction timescales in the solar atmospheric plasma. To this end, in this Thesis we have carried out an exhaustive comparative study [2] between the CO abundances obtained assuming instantaneous chemical equilibrium and that obtained by following the chemical evolution in one-dimensional hydrodynamical models of the solar atmosphere. These models show the generation of acoustic wave trains which propagate upwards in the atmosphere until transforming into shocks [15]. As shown in Fig. 2, our results indicate that the CO line radiation observed close to the edge of the solar disk comes from atmospheric heights not greater than ∼700 km. Above this height, the CO abundance given by the ICE approximation leads to an overestimation of more than two orders of magnitude. It is of interest to point out that our chemical evolution codes can also be used to investigate the formation processes of complex molecules. The exact chemical mechanisms which produce such molecules with more than 10 atoms is not correctly known. Actually, very complex molecules are found in the interstellar medium. One of the most striking has been the detection of benzene C6 H6 in circumstellar envelopes [16]. Radiative Transfer in Molecular Lines. Astrophysical Applications 111 2.2 The Zeeman and Hanle Effects in Molecular Lines The study of atomic and/or molecular lines allows us to obtain information about the physical properties of the medium the radiation is coming from. In addition to its intensity and frequency, light is characterized by its state of polarization. Spectropolarimetry provides an incredible amount of information about phenomena in which a break of the spherical symmetry occurs in the astrophysical system in which the spectral lines are formed. We may have stellar geometrical asymmetries induced by the presence of another companion star in the case of a binary star, global anisotropies in the radiation field of non-resolved stars, local anisotropies in the radiation field, the presence of magnetic fields, etc. All these phenomena produce recognizable signatures in the polarization state of the observed light. If the study of such polarization phenomena is tackled within the framework of the quantum theory of polarization [27], we can obtain reliable information about the physical properties in a variety of astrophysical objects The polarization state of the light can be quantified by using the Stokes parameters [13]. The light intensity is represented by I, Q is the intensity difference between vertical and horizontal linear polarization, U the intensity difference between linear polarization at +45◦ and −45◦ , while V is the intensity difference between right-handed and left-handed circular polarization. Observationally, Stokes parameters can be easily obtained when working on the radio spectral domain since the detectors are directly sensitive to the polarization state of the light. For the case of shorter wavelengths, the technique is much more complicated and it relies on modulation schemes [25]. In the special case of solar physics, the present instrumentation is very sophisticated. Nowadays, very sensitive polarimeters based on several modulation schemes have been developed which allow us to measure the Stokes parameters from the infrared to the ultraviolet (TIP, ASP, ZIMPOL, etc.). Such sensitive polarimeters allow the detailed investigation of physical processes which produce very weak signatures in the polarization state of the light (see the proceedings of the Solar Polarization 3 workshop edited by Trujillo Bueno & Sánchez Almeida). In this way, the topology and strength of solar magnetic fields can be inferred via the physical interpretation of spectropolarimetric observations. The application of spectropolarimetry to night-time astronomy is currently in expansion thanks to the construction of several polarimeters: the MuSiCoS6 échelle spectro-polarimeter for the 2-m Bernard Lyot Télescope at Pic du Midi, the Semel’s visitor polarimeter on the UCL Echelle Spectrograph of the 3.9-m Anglo-Australian Telescope or the ESPaDOnS7 polarimeter mounted on the Canada-France-Hawaii Telescope. These polarimeters are being used or will be used for the investigation of magnetic fields in magnetic stars (see the proceedings of the above-mentioned Solar Polarization 3 workshop 2003). 6 7 http://www.ast.obs-mip.fr/projets/espadons/polarmus.html http://www.ast.obs-mip.fr/projets/espadons/espadons.html 112 A. Asensio Ramos An important part of this Thesis focused on the investigation of polarization signals produced by molecular lines. Similar to what happens with the atomic case, the coupling between the different angular momenta in the molecule with the external magnetic field produces a magnetic moment and a precession of the total angular momentum around the magnetic field vector. This is the molecular Zeeman effect. The influence of the magnetic field on the level structure of the molecules produces an observable effect on the polarization state of the light emitted or absorbed by the molecule. The investigation of the Zeeman effect in molecules is much more complex than in atoms, even for the simplest coupling cases. The main reason for this complexity is the presence of rotation in the molecule, which produces additional angular momentum couplings[23]. There are formulae applicable for the calculation of the splittings and strengths of the Zeeman components for electronic doublet states [31]. Recently, this formulation has been extended to electronic states of arbitrary multiplicity. In this Thesis, we have developed a formalism (and an ensuing computer code) that allows us to obtain the splittings and strengths of the Zeeman components for any transition between two arbitrary rotational levels of arbitrary electronic states in a diatomic molecule [8]. Our approach is capable of treating the inclusion of any refinement in the quantum description of the molecular motion by only adding the corresponding contribution to the total molecular hamiltonian. The calculation of the hamiltonian matrix elements is carried out by taking advantage of the powerful tools of Racah algebra [14]. The observation of spectral line polarization in cold and magnetized regions on the solar surface (sunspots) gives us information about the magnetic field in these regions. We have shown that molecules usually present anomalous polarization profiles that can be explained via the different strengths of the angular momenta couplings. Some of the rotational levels of several electronic states in diatomic molecules present strong interactions with nearby levels. This produces a transition from the Zeeman regime to the PaschenBack regime, often for rather low magnetic field strengths. Some spectral lines of molecular species present Stokes profiles with strongly anomalous behaviors. This is the case of the CN lines we have observed with TIP in the near infrared [5], as shown in Fig. 3. Apart from their diagnostic capabilities in solar physics, the molecular Zeeman effect has recently started to be applied for obtaining information about magnetic fields in stars with very strong magnetic fields. Such strong fields are thought to produce very cold spots in the surface of the stars and molecules constitute one of the few observable tracers of the physical conditions in these regions[36]. A serious problem is the blending of molecular lines with important atomic lines used for diagnostic purposes. We know that molecular bands are characterized by a huge amount of lines produced by their rotational structure. Therefore, it is quite probable that a molecular line is blended with an atomic line. It turns out necessary to include a huge number of both atomic and molecular lines in our spectral synthesis codes to obtain correct information about Radiative Transfer in Molecular Lines. Astrophysical Applications 113 Fig. 3. Observed profiles (dotted line) and those resulting from a model obtained from the inversion of the observed Stokes profiles of CN and OH (solid line). Note the appearance of strongly distorted CN linear polarization profiles (antisymmetric when they are normally symmetric in the Zeeman regime). The energy levels of these transitions are in the intermediate Paschen-Back regime for typical umbral magnetic fields. The Stokes profiles are normalized to the continuum intensity calculated in the Harvard-Smithsonian Reference Atmosphere (HSRA). the physical properties of the atmosphere when comparing with observations. If the molecular line is magnetically sensitive, the influence of the blend on the polarization state has to be included in the forward modeling. This analysis of the physical properties of the stellar atmosphere is usually accomplished with the aid of inversion codes. We have developed an LTE inversion code which allows to include atomic and/or molecular lines in the forward modeling. An application of this code is shown in Fig. 3, where we have used the OH and CN lines that we have observed in the umbra of a sunspot [4] to obtain information about its thermodynamic and magnetic structure. Ideally, one should include a suitable set of spectral lines which can trace the physical properties at different heights. An extra novel subject we have considered in this Thesis is that of scattering polarization in molecular lines [3]. We present a very detailed and systematic investigation of the polarization signals produced by scattering processes in several molecular lines. When one observes the second solar spectrum of the Sun (the linearly polarized spectrum close to the solar limb) many conspicuous signals corresponding to molecular lines appear [22]. They are generated by the scattering of radiation. Such polarization signals are modified by the presence 114 A. Asensio Ramos Fig. 4. Number density of MgH and C2 relative to the maximum concentration in the 3D snapshot model at heights of 145.5 km and 245.5 km. Note that h=0 km is, as usual, the height where we have optical depth unity in the continuum at 5000 Å for vertical incidence. Note also that the molecular abundance is higher in the upflowing material than in the downflowing plasma. of a magnetic field, which is known as the Hanle effect. The observed linear polarization of many atomic species present variations across the solar cycle. On the contrary, molecular lines do not show any variation during the solar cycle. The scattering polarization of molecular lines are indeed sensitive to the presence of a magnetic field via the Hanle effect, with critical fields that are in the same range as for atomic lines [26, 34]. We show in this Thesis that this sensitivity to the Hanle effect is found for fields of the order of 10 G, and that the apparent insensitivity of the molecular lines has to be assigned to another physical effect. In this direction, we have demonstrated that this behavior may be explained when taking into account that the molecules which generate the linear polarization signals are formed in a three-dimensional medium. When obtaining the molecular abundances in such 3D models, we find that they are larger in the upflowing material than in the downflowing plasma, at least in the regions where the molecular lines are “formed” [35]. This explains the apparent insensitivity of the molecular lines to the magnetic field. Additionally, we show how to obtain information about the distribution of weak magnetic fields in the “quiet” solar photosphere. 3 Work after the Thesis The work carried out during the Thesis with the development of innovative diagnostic tools is facilitating the investigation of new problems in Astrophysics. In the field of the numerical solution of the non-LTE problem, we have extended the escape probability approximate formalism to an exact method [20]. This method is of great interest because it can be integrated in existing numerical codes based on the escape probability method with a reduced set of Radiative Transfer in Molecular Lines. Astrophysical Applications 115 changes. We have built a numerical code that we plan to include in the existing packages for the analysis of observations of the future Herschel telescope. We have also applied our experience and knowledge of the theory of atomic polarization in spectral lines [27] to the investigation of dichroic SiO masers induced by the radiation anisotropy, with emphasis on clarifying the influence of a magnetic field on the SiO polarization properties [6]. We have demonstrated that masers can appear even in the absence of a global population inversion in the rotational levels of SiO. Dichroic masers appear due to a population inversion produced by the population imbalances between the magnetic sublevels of the rotational levels, which are induced by an anisotropic radiation field. We have also investigated in detail the effect of collisions and magnetic fields on the depolarization of the MgH lines observed close to the solar limb [7]. We have shown that collisions seem to be very efficient in depolarizing the rotational levels of MgH. In this case, the strength of the magnetic field in the upflowing regions of the “quiet” solar photosphere cannot be much larger than 10 G, which reinforces our conclusion that there is a vast amount of hidden magnetic energy in the downflowing regions [35]. Acknowledgements: I would like to express my most sincere gratitude to my supervisor, Javier Trujillo Bueno, for his guidance and encouragement at all stages of my work and for showing me the exciting world of spectropolarimetry. I also thank my co-supervisor, José Cernicharo Quintanilla, for pointing out several key problems in Molecular Astrophysics. This research has been partly funded by the Ministerio de Educación y Ciencia through project AYA2004-05792. References 1. Asensio Ramos, A., Trujillo Bueno, J., & Cernicharo, J.: in The Promise of the Herschel Space Observatory, ed. G. L. Pilbratt, J. Cernicharo, A. M. Heras, T. Prusti, & R. Harris, ESA-SP 460, San Francisco, 265 (2001) 2. Asensio Ramos, A., Trujillo Bueno, J., Carlsson, M., & Cernicharo, J.: ApJ, 588, L61 (2003) 3. Asensio Ramos, A. & Trujillo Bueno, J.: in Solar Polarization 3, eds. J. Trujillo Bueno & J. Sánchez Almeida, ASP Conf. Ser. Vol 307, 195 (2003) 4. Asensio Ramos, A., Trujillo Bueno, J., Bianda, M., Manso Sainz, R., & Uitenbroek, H.: ApJ, 611, L61 (2004) 5. Asensio Ramos, A., Trujillo Bueno, J. & Collados, M.: ApJ, 623, L57 (2005) 6. Asensio Ramos, A., Landi Degl’Innocenti, E., & Trujillo Bueno, J.: ApJ, 625, 985 (2005) 7. Asensio Ramos, A. & Trujillo Bueno, J.: ApJ, 635, L109 (2006) 8. Asensio Ramos, A. & Trujillo Bueno, J.: ApJ, 636, 548 (2006) 9. Asensio Ramos, A. & Trujillo Bueno, J.: In 3rd GRETA meeting, ed. Ph. Stee, EAS Publications Series, 18, 25 (2006) 10. Asplund, M., & Garcı́a Pérez, A. E.: A&A, 372, 601 (2001) 116 A. Asensio Ramos 11. Asplund, M., Grevesse, N., Sauval, A. J., Allende Prieto, C., & Kiselman, D.: A&A, 417, 751 (2004) 12. Ayres, T. R.: ApJ, 244, 1064 (1981) 13. Born, M., & Wolf, E.: Principles of Optics: Electromagnetic Theory of Propagation of Light (Oxford: Pergamon Press) (1980) 14. Brown, J. M., & Carrington, A.: Rotational Spectroscopy of Diatomic Molecules (Cambridge: Cambridge University Press) (2003) 15. Carlsson, M., & Stein, R. F.: ApJ, 440, L29 (1995) 16. Cernicharo, J., Heras, A. M., Tielens, A. G. G. M., Pardo, J. R., Herpin, F., Guélin, M., & Waters, L. B. F. M.: ApJ, 546, L123 (2001) 17. Cernicharo, J., Goicoechea, J. R., Pardo, J. R., & Asensio Ramos, A.: ApJ, 642, 940 (2006) 18. Dunham, T. J., & Adams, W. S.: Publ. Am. Astr. Soc., 9, 5 (1937) 19. Dunham, T. J., & Adams, W. S.: PASP, 49, 26 (1937) 20. Elitzur, M., & Asensio Ramos, A.: MNRAS, 365, 779 (2006) 21. Feuchtgruber, H., Helmich, F. P., van Dishoeck, E. F., & Wright, C. M. : ApJ, 535, L111 (2000) 22. Gandorfer, A.: The Second Solar Spectrum, Vol. I: 4625 Å to 6995 Å (Zurich: vdf) (2000) 23. Herzberg, G.: Molecular Spectra and Molecular Structure. I. Spectra of Diatomic Molecules (New York: Van Nostrand Company) (1950) 24. Höflich, P.: in Stellar Atmosphere Modeling, ed. I. Hubeny, D. Mihalas, & K. Werner, ASP Conf. Ser. 288 (San Francisco: ASP), 371 (2003) 25. Keller, C.: in Astrophysical Spectropolarimetry, ed. J. Trujillo Bueno, F. Moreno-Insertis, & F. Sánchez, Proceedings of the XII Canary Islands Winter School of Astrophysics (Cambridge, UK: Cambridge University Press), 303 (2002) 26. Landi Degl’Innocenti, E.: in Solar Polarization 3, ed. J. Trujillo Bueno & J. Sánchez Almeida, ASP Conf. Ser. Vol 307, 164 (2003) 27. Landi Degl’Innocenti, E., & Landolfi, M.: Polarization in Spectral Lines, (Kluwer Academic Publishers 2004) (2004) 28. Noyes, R. W., & Hall, D. N. B.: BAAS, 4, 389 (1972) 29. Penzias, A. A., & Wilson, R. W.: ApJ, 142, 419 (1965) 30. Rybicki, G., B., & Hummer, D. G.: A&A, 249, 720 (1991) 31. Schadee, A.: J. Quant. Spectrosc. Radiat. Transfer, 19, 517 (1978) 32. Shchukina, N., & Trujillo Bueno, J.: ApJ, 550, 970 (2001) 33. Trujillo Bueno, J., & Fabiani Bendicho, P.: ApJ, 455, 646 (1995) 34. Trujillo Bueno, J.: in Solar Polarization 3, ed. J. Trujillo Bueno & J. Sánchez Almeida, ASP Conf. Ser. Vol 307, 407 (2003) 35. Trujillo Bueno, J., Shchukina, N., & Asensio Ramos, A.: Nature, 430, 326 (2004) 36. Valenti, J. A., Johns-Krull, C. M., & Piskunov, N.: in ASP Con. Ser., Vol. 223, The 11th Cool Stars, Stellar Systems and the Sun, ed. R. J. Garcı́a López, R. R., & M. R. Zapatero Osorio, 1579 (2001) 37. van Zadelhoff, G. J., Dullemond, C. P., van der Tak, F. F. S., Yates, J. A., Doty, S. D., Ossenkopf, V., Hogerheijde, M. R., Juvela, M., Wiesemeyer, H., & Schöier, F.: A&A, 395, 373 (2002) 38. Wallace, L., Bernath, P., Livingston, W., Hinkle, K., Busler, J., Guo, B., & Zhang, K.: Science, 268, 1155 (1995) 39. Wilson, R. W., Jefferts, K. B., & Penzias, A. A. 1970, ApJ, 161, L43 (1970) The star formation history of early-type galaxies as a function of environment P. Sánchez-Blázquez University of Central Lancashire, psanchez-blazquez@uclan.ac.uk Summary. I present a short summary of some of the results from my thesis. In this work we analyse a sample of 98 early-type galaxies situated in different environments. Using new models, that include a new and improved stellar library, we derive mean ages, metallicities and study the qualitative behaviour of different chemical elements. We conclude that more massive galaxies formed their stars on shorter time-scales than less massive ones. The formation epoch and time-scale of the star formation in the most massive galaxies is very similar in the field and in more dense environments, but the star formation is progressively more extended for less massive galaxies in less dense environments. 1 Introduction There have long been two competing views on the star formation history of early-type galaxies in the present day Universe. The modern version of the classical monolithic collapse scenario puts the stress on elliptical assembly out of gaseous material (that is, with dissipation), in the form of either a unique cloud or many gaseous clumps, but not out of preexisting stars. In this scenario, the stars form at high z and on short time-scales relative to spiral galaxies. The competing hierarchical scenario propounds that galaxies form through successive, non-dissipative, random mergers of subunits over a wide redshift range. The first scenario succeed in explaining the tight relations followed by the elliptical family, such as the Fundamental Plane, the colour– magnitude, and the Mg2 –σ relationships (e.g. [11, 2, 18]). However, detailed comparison between spectral characteristic in the central parts of early-type galaxies and stellar population models have reveal a large dispersion in the ages [16, 37, 5], a large percentage of those show kinematical and dynamical peculiarities (e.g. [10]), as well as presence of shells and ripples, indicative of recent interactions [32]. A natural outcome of the hierarchical scenarios is that haloes in regions of the Universe that are destined to form a cluster collapse earlier and merge more rapidly (e.g. [19, 9]). Therefore, the study of the stellar 118 P. Sánchez-Blázquez content of early-type galaxies in different environments should be a good test for the hierarchical scenarios of galaxy formation. 2 Observations and sample We analyse a sample of 98 early-type galaxies, which includes ellipticals (E) and lenticular (S0) spanning a large range in velocity dispersion (from 40 km s−1 to 400 km s−1 ). As one of the main goals of this work is to study the influence of the environment on the star formation history of early-type galaxies, the sample contains galaxies in the field, poor groups, and in the Virgo, Coma, and some Abell galaxy clusters. For the purpose of this work, we have divided the sample in two main groups that we call hereafter high density environment galaxies (HDEGs) and low density environment galaxies (LDEGs). Long-slit spectroscopy was carried out in four observing runs using two different telescopes. Typical signal-to-noise ratios per Å, measured in the range between 3500 and 6500 Å, are 110 and 50 for the LDEGs and HDEGs galaxies, respectively. The wavelength coverage varies between different runs, but all includes the range between 3500 and 5250 Å, which allows us to the measure the D4000 break and 15 Lick/IDS indices (from HδA to Mgb). Previous works have used Lick/IDS line-strength indices to derive mean ages and metallicities using evolutionary synthesis models. Here we follow a similar approach, deriving the SSP parameters (age and metallicity) by comparing the observed line-strengths with the predicted index–index diagrams from a new set of models by Vazdekis et al. (2006, in preparation, V06 hereafter). These models are an updated version of those described by [40], improved by the inclusion of a new stellar library (MILES) recently observed by [31]. This library contains 985 stars, carefully selected to cover the atmospheric parameter space in an homogeneous way. In particular, the library span a range of metallicities from [Fe/H]∼ −2.7 to +1.0 and a wide range of effective temperatures. The inclusion of this library reduces the uncertainties in the models, especially at metallicities departing from solar. Since the stars of the library are relatively flux calibrated, these models are able to predict not only individual features for a population of a given age and metallicity, but also the whole spectral energy distribution (SED). This allows us to analyse the spectra of the galaxies at their own resolution, given by their internal velocity and instrumental broadening (see e.g. [41]). The synthetic spectra have a spectral resolution of 2.4 Å and cover the spectral range 3500-7500Å. To quantify the age and metallicity values, we interpolated in the grids using bivariate polynomials, as described in [6]. [26] first notice that, when using Mg2 to derive metallicities, those were much larger than when using Fe5270. Several works have confirmed this result and is normally attributed to the fact that early-type galaxies have different chemical composition than the one found in the solar neighborhood, (which is the partition we use to build the models). For example, the metallicities The star formation history of early-type galaxies 119 measured in giants early-type galaxies using indices such as CN2 , Mgb, and C4668 are larger than the metallicities measured with Fe-sensitive indices such as Fe4383, and that have been commonly interpreted as an overabundance of Mg, C and N with respect to Fe in these systems. Although the presence of non-solar abundances ratios is one of the major drawbacks to derive absolute ages and metallicities, their study provide us with invaluable information to compare with chemical evolutionary models and, in such a way, derive star formation histories, avoiding the problems due to the age-metallicity degeneracy. Unfortunately, the derivation of the detailed abundances through the comparison with stellar population models is still in its infancy. However, the relative trends between different elements and other properties of the galaxies can still provide us with very useful information. For the rest of the analysis, we make the assumption that the differences between the metallicities derived from various index–index diagrams, combining Hβ with other metallicity indicators, are due to changes in the sensitivity of these indicators to variations of different chemical abundances. Figure 1 shows the ages and metallicities obtained in different index–index diagrams as a function of the velocity dispersion for both HDEGs (filled symbols) and LDEGs (open symbols). As can be seen, the relation between the metallicity and the velocity dispersion depends on the index used to derive this parameter. The relation between σ and the metallicity also depends on the environment. For HDEGs, there is a relation between the metallicity and the velocity dispersion, no matter which index is used to measured metallicity. However, for LDEGs, the mass metallicity is only evident when Mgb (specially sensitive to Mg abundance) is used. However, the relationship is much flatter when Fe4383 (sensitive to Fe abundances), CN2 (sensitive to N abundances) or C4668 (sensitive to C abundances) are used. We can interpret these differences in terms of differences in the star formation histories of these galaxies, as the different elements are release to the interstellar medium in different time-scales. As can be seen in the Figures, the most massive galaxies of the sample show a very similar behaviour in both environments, having the highest ratio of Mg/Fe1 . This indicates that giants early-type galaxies formed their stars in very short timescales, and that the environment did not affect much the duration of the star formation. However, less massive LDEGs show a increase abundance of Fe, C and N compared with their counterparts in dense clusters, indicating that their star formation history (SFH) have been more extended. The bottom panel of the Fig. 1 show the relation between the age and the velocity dispersion. It can be seen that, while the relation is flat for the HDEGs, a correlation exists between the age and velocity dispersion for the LDEGs, in the sense that low velocity dispersion galaxies tend to be younger. This is in agreement with the suggestion by [36], who found differences in the (σ, t) plane between galaxies in the field and in the Fornax cluster. Interest1 Note that we are not measuring real abundances but qualitative trends 120 P. Sánchez-Blázquez ingly, [17] did not find any correlation between age and velocity dispersion in her study of a sample of galaxies in the Coma cluster, although she found a considerable dispersion in the ages of the galaxies. [5] also found younger ages for lower σ galaxies in a sample of Virgo galaxies and galaxies in lower environments. [35] did not find a significant trend between the age and the velocity dispersion in either their sample of high- or low-density environment galaxies. However, they argued that the correlated errors of age and metallicity tend to dilute a correlation between age and the velocity dispersion and that their observational data are best reproduced by a relatively flat, but significant correlation. It is also clear from Fig. 1 that the age dispersion for less massive galaxies is higher than for the more massive ones, in agreement with other studies. Fig. 1. Relations between metallicities, obtained with different indicators, and age against velocity dispersion for the sample of galaxies. Open symbols represent galaxies in low-density environments (LDEGs), while filled symbols indicate galaxies in high density environments (HDEGs). Squares correspond to S0 galaxies, while elliptical galaxies are represented with circles. Grey and black lines show the linear fit, weighting with the errors in both axes, to the LDEGs and HDEGs respectively. The star formation history of early-type galaxies 121 2.1 Simple stellar population or multiple burst? It has been seen above that the LDEGs span a broad range in their apparent mean ages and that, in some cases, these ages are very low. Since the models assume a unique burst of star formation, these low values can indicate either that these galaxies are genuinely young, i.e., they formed all their stars recently, or that most of their stars were formed at early epochs, but that they have undergone later episodes of star formation involving a certain percentage of the total mass of the galaxy (see [36]). In the latter case, the apparent mean age would depend on the relative light contributions of the different components to the considered spectral range. To distinguish between the two scenarios, we carried out a comparison of the observed galaxy spectra with the synthetic spectra extracted from the V06 models in two different wavelength ranges: 3650–4050 Å and 4750–5150 Å. Figure 2 show the age distributions obtained in both wavelength ranges for LDEG and HDEG, respectively. It is apparent from Fig. 2 that, for LDEGs, there is a significant difference between the ages obtained in the two different regions of the spectra. This is difficult to understand if all the stars were formed in a single burst, and it suggests that many LDEGs are composite systems consisting of an underlying old population plus, at least, a later star formation burst. Fig. 2. Distribution of ages obtained comparing the synthetic spectra from V06 with the spectral energy distribution of LDEGs and HDEGs. The empty histogram shows the ages obtained with the comparison in the spectral range 4750–5150 Å, while the shaded histogram the ages obtained comparing the region from 3650–4050 Å. To study this in more detail, we have built different composite spectra in which we have added different components of metallicity [M/H]=+0.2 and ages ranging from 2.51 to 14.12 Gyr to an old population of 15.85 Gyr and metallicity [M/H]=−0.38 dex. The percentages of these two components were chosen to be 70 and 30% (model 1, solid line), 80 and 20% (model 2, dashed line), and 90 and 10% (model 3, dotted line) in mass, respectively. Figure 3 shows the relation between the derived ages in these two spectral ranges for different models and, over-plotted, the derived ages for the LDEGs (crosses). As can be seen, although is difficult to match the observed points with single 122 P. Sánchez-Blázquez scenarios, as the contribution in mass and the look-back time of the star formation event are highly degenerate, the combination of an old population and a burst of star formation would lead to similar trends in the derived ages as the observed for the LDEGs. We also note that, to reproduce the observed trends, the metallicity of the young component must be higher than the metallicity of the underlying old population, in agreement with the findings of other authors (e.g. [15, 37, 35]). We note here that the differences between the ages derived in the two spectral ranges do not follow a simple relation. The shape of this relation depends on the difference in the light contribution of the burst to the considered wavelength regions. The difference in the light fraction of the burst between the two considered spectral regions increase with the age of the burst and that is the reason why the differences between the ages calculated in two different spectral ranges increase also with this parameter. On the other Fig. 3. Comparison of the ages derived in two different spectral ranges using the models of V06. The asterisks are the values calculated for the LDEG spectra, while the filled circles are the ages derived from the composite models in which two populations of different ages and metallicities are added (see text for details). The solid, dashed, and dotted lines connect the various two-component model combinations for the 30:70, 20:80, and 10:90 young:old population mass ratios, respectively. The age of the younger component increases from lower left to upper right of the diagram. hand, the distribution of ages for HDEGs (righ panel of Fig. 2) shows no such clear dichotomy (see mean values in the insets). This is compatible with the idea that these galaxies constitute a more homogeneous (coeval) sample that have undergone their last episode of star formation at higher redshift. This interpretation is in agreement with our findings in [28] and [30]. The star formation history of early-type galaxies 123 3 Age-metallicity relation Several authors have noted that when the age and metallicity obtained from an index–index diagram are plotted together, they show a correlation in the sense that younger galaxies seem to be also more metal rich (e.g. [38, 36, 34]). This relation is expected if, during their evolution, galaxies have undergone several episodes of star formation (or have had an extended star formation history), in which the new stars formed from pre-enriched gas by the previous generations of stars. Furthermore, the existence of an age–metallicity relation has implications in the interpretation of the scale-relations. The low dispersion in the Mg2 –σ or the color–magnitude relations and the existence of a fundamental plane have been common arguments in favour of the hypothesis that elliptical galaxies are old systems that formed all their stars at high redshift and evolved passively since then ([2, 3]). However, some authors, e.g. [38, 36, 15, 17] have studied the scale relationships showing that a possible age–metallicity degeneracy would constitute a conspiracy to preserve the low dispersion in those relationships even when a relatively large fraction of galaxies contain young stars. When plotting the age and metallicity derived from an index-index diagram is clear that younger galaxies appear to be also more metal rich. However, when age and metallicity are measured in a partially degenerated index– index diagram, the correlation of the errors in both parameters tends to create an artificial anti-correlation between them [22], so it is difficult to disentangle whether the relation is real or an artifact due to the age-metallicity degeneracy. A very simple way to check if the age-metallicity relationship is not due solely to this effect is to compare the age and the metallicity obtained in two completely independent diagrams. Figure 4 shows the age–metallicity relation where the ages have been measured in a Hβ–Fe4531 diagram and the metallicities in a HδF –Fe4383 diagram for LDEGs (left panel) and HDEGS (right panel). A non-parametric Spearman rank-order correlation test gives a correlation coefficient of −0.47 corresponding to a significance level of 0.0002 for LDEGs. Although the slope of the relation is flatter (−0.237 ± 0.076) than the one obtained by measuring the ages and metallicities in a Fe4383–Hβ diagram, there is still a significant correlation, which confirms that the age-metallicity relation is not entirely due to the correlation of the errors in both parameters for galaxies in low density environments. For HDEGs we do not find any correlation between both parameters (the non-parametric Spearman rank order coefficient is 0.039 with a significance level of 0.422). 4 Discussion The results presented here indicate that HDEGs constitute a more homogeneous family than LDEGs; their stellar populations can be explained under 124 P. Sánchez-Blázquez Fig. 4. Comparison of the ages and metallicities obtained from completely independent diagrams for the LDEGs (left panel) and HDEGs (right panel). The line represents a least-square fit, minimizing the residuals in both directions, x and y. the hypothesis of a single population, and they are, on average, older. In Fig. 1, it has been shown that this sub-sample of galaxies exhibits a relation between the metallicity and the velocity dispersion, no matter which indices are used to derive this parameter, but, on the contrary, there is no age variation with velocity dispersion. For LDEGs, however, the age dispersion is higher and their populations are best explained as a composition of different bursts of star formation. The hierarchical clustering models of structure formation predict different star formation histories for galaxies situated in different environments. In these models, clusters of galaxies are formed from the highest peaks in the primordial density fluctuations. It is there where the merging of dark matter haloes, which contained the first galaxies, leads to galaxies dominated by a bulge at high redshifts (z ≥ 2). The mergers of galaxies and the acquisition of cold gas cannot continue once the relative velocity dispersion between galaxies is higher than 500 km s−1 , which makes the occurrence of further star formation episodes in these galaxies more difficult. This truncated star formation history also explains the higher [Mg/Fe] found in HDEGs with respect to the values in younger looking LDEGs [29]. On the other hand, the star formation in LDEGs has probably extended over a longer period of time, due to the occurrence of more star formation events or due to a longer single episode of star formation. This scenario was proposed to explain the differences between N, and maybe C, when comparing galaxies in different environments [28] We speculate that LDEGs and HDEGs could have initially presented similar relations between the metallicity and the velocity dispersion after their first massive star formation episode. However, if LDEGs have suffered subsequent episodes of star formation, the original correlation between metallicity and potential well (or mass) could have been erased, since other processes could have also played a role in defining the final metal content of the galaxies. The new stars, formed in the more recent events, would do it from a gas more enriched in the elements produced by low- and intermediate-mass stars, due to the higher active evolution timescale of these galaxies. If these star formation processes have had a greater relative influence (a larger ratio between the burst strength and the total The star formation history of early-type galaxies 125 galaxy mass) in less massive galaxies, as suggested by the age–σ relation, this would destroy the original relation between mass and metallicity (increasing Fe, C, and N in low velocity dispersion galaxies). Furthermore, this would result in a relation between age and metallicity (as inferred from Fe features) in LDEGs, but not in HDEGs, as it is found in here. Another possibility is that less massive galaxies have actually experienced a more extended star formation history than more massive galaxies. This latter possibility is favoured by some recent studies that found a depletion in the luminosity function of red galaxies towards the faint end [33, 8]. Other authors have found differences between Fig. 5. Age–metallicity relation for the sample of low-density environment galaxies when these parameters are measured in a Mgb–Hβ diagram. The line indicate a least-square fit to the data, minimizing the residuals in both directions x and y. the mass–metallicity relation of galaxies in different environments. [36] found that there is a velocity dispersion–metallicity relation for old cluster galaxies, but no comparable relation exists for field ellipticals. This result is compatible with ours, with the difference that we still find a steep relation between the metallicity and the velocity dispersion for LDEGs when the metallicity is measured with Mgb. Actually, [36] also found a relation between what they called the enhanced elements (including Mg) and velocity dispersion for all the galaxies in their sample. If, as we have argued, the age–metallicity relation is a consequence of later episodes of star formation, and the relative enrichment has been more pronounced in the Fe-peak elements, we would expect differences in the agemetallicity relation when the metallicity is measured using an index with a different sensitivity to changes in Fe and Mg. Figure 5 shows the age–metallicity relation when these parameters are measured in a Mgb–Hβ diagram. The non-parametric rank order coefficient is 0.177, with a significance level of 0.10. Certainly, there is not a significant correlation between these two parameters when the Mgb index is used instead of Fe4383. We need to stress again that we are not calculating chemical abundances in this paper. The metallicity measured with Mgb does not correspond to the abundance of Mg, nor does the metallicity measured with Fe4383 correspond to a Fe abundance. 126 P. Sánchez-Blázquez We argue though, that the different behaviours of the metallicities calculated with different indices are the consequence of their different sensitivities to the variation of different chemical species. In this specific case, the flatter slope of the age-metallicity relation when a more (less) sensitive Mg (Fe) index is used is in agreement with our scenario. Interestingly, the more massive galaxies in low density environments show a behaviour very similar to the massive galaxies of the Coma cluster. These very massive galaxies tend to have boxy isophotes, which can be explained by models of mergers without gas [4, 25, 14], since a few percent of the mass in gas is sufficient to destroy boxy orbits and impart high global rotation [1]. Furthermore, boxy galaxies tend to have flat inner profiles [14]. N-body simulations of merging galaxies with central black holes [12, 27, 24] show that cores can indeed form in such merger remnants. Recently, [23] have found that power-law galaxies, on average, have steeper colour gradients than do core galaxies (although the difference is small). This result is compatible with the idea that power-law galaxies have formed in gas-rich mergers while core galaxies have formed from free-gas mergers, which would cause a dilution in the metallicity gradient. Actually, these mergers without gas have been observed in clusters at z=0.8 [39]. The existence of these gas-free mergers indicates that the epoch of assembly does not necessarily coincide with the epoch of formation of the bulk of stars. This scenario could bring the hierarchical models of galaxy formation into agreement with the observed trends of age with mass for elliptical galaxies in LDEGs. These trends (low-σ galaxies appearing to be younger) are completely opposite to what is expected under these scenarios of galaxy formation, which predict that larger galaxies assemble at later times than small ones (e.g. [19]). However, these predictions are made under the assumption that all the gas cooled off and formed stars when the haloes were assembled. However, other processes, such as supernova feedback, may play a role in regulating the rate at which stars form in these systems (e.g. [21]). Several mechanisms have been proposed to explain the appearance that low-mass galaxies have suffered a more extended star formation history. [20] suggests that UV background radiation is a possible candidate because it suppresses cooling and star formation more strongly in lower mass systems [13], and is expected to extend the duration of star formation. [7] have recently built N-body-tree-SPH simulations incorporating cooling, star formation, energy feedback, and chemical evolution. These authors find that the star formation history is governed by the initial density and total mass of the galaxy, and that the interplay of the above processes results in a more extended star formation history in low-mass galaxies. Until we understand the role of these mechanisms completely, we will not be able to rule out different processes of galaxy formation. The star formation history of early-type galaxies 127 References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. J. E. Barnes, L. Hernquist: ApJ, 471, 115 (1996) R. Bender, D. Burstein, S. M. Faber: ApJ, 411, 153 (1993) M. Bernardi, A. Renzini, L.N. da Costa, et al.: ApJ, 508, 143 (1998) J. Binney, M. Petrou: MNRAS, 214, 449 (1985) N. Caldwell, J. A. Rose, K. D. Concannon: AJ, 125, 2891 (2003) N. Cardiel, J. Gorgas, P. Sánchez-Blázquez, et al.: A&A, 409, 511 (2003) C. Chiosi, G. Carraro: MNRAS, 335, 335 (2002) G. De Lucia, B. M. Poggianti, A. Aragón-Salamanca et al.: ApJL: 610, L77 (2004) G. De Lucia, V. Springel, S. D. M. White, D. Croton, & G. Kauffmann: MNRAS, 366, 499 (2006) P. T. de Zeeuw, M. Bureau, E. Emsellem, et al.: MNRAS, 329, 513 (2002) S. Djorgovski, M. Davis: ApJ, 313, 59 (1987) T. Ebisuzaki, J. Makino, S. K. Okumura: Nature, 354, 212 (1991) G. Efstathiou: MNRAS, 256, 43P (1992) S. M. Faber, S. Tremaine, E. A. Ajhar, et al.: AJ, 114, 1771 (1997) I. Ferreras, S. Charlot, J. Silk: ApJ, 521, 81 (1999) J. J. González: Ph.D. Thesis (1993) I. Jørgensen: MNRAS, 306, 607 (1999) I. Jørgensen, M. Franx, J. Hjorth, P. G. van Dokkum: MNRAS, 308, 833 (1999) G. Kauffmann, S. Charlot: MNRAS, 294, 705 (1998) D. Kawata: ApJ, 558, 598 (2001) D. Kawata, B. K. Gibson: MNRAS, 346, 135 (2003) H. Kuntschner, J. R. Lucey, R. J. Smith, et al.: MNRAS, 323, 615 (2001) T. R. Lauer, S. M. Faber, K. Gebhardt et al.: AJ, 129, 2138 (2005) M. Milosavljević, D. Merritt:ApJ, 563, 34 (2001) J.-L. Nieto, R. Bender, P. Surma: A&A, 244, L37 (1991) R. F. Peletier: Ph.D. Thesis (1989) G. D. Quinlan, L. Hernquist: New Astronomy, 2, 533 (1997) P. Sánchez-Blázquez, J. Gorgas, et al.: ApJL, 590, L91 (2003) P. Sánchez-Blázquez, J. Gorgas, N. Cardiel, J. J. González : A&A, 457, 787 (2006) —. : A&A, 457, 809 (2006) P. Sánchez-Blázquez, R. F. Peletier, J. Jiménez-Vicente et al. : MNRAS, 371, 703 (2006) F. Schweizer, P. Seitzer, S. M. Faber, et al.: ApJL, 364, L33 (1990) I. Smail, H. Kuntschner, T. Kodama, et al.: MNRAS, 323, 839 (2001) A. I. Terlevich, D. A. Forbes: MNRAS, 330, 547 (2002) D. Thomas, C. Maraston, R. Bender, C. Mendes de Oliveira: ApJ, 621, 673 (2005) S. C. Trager, S. M. Faber, G. Worthey, J. J. González: AJ, 120, 165 (2000) —. : AJ, 119, 1645 (2000) S. C. Trager, G. Worthey, S. M. Faber, et al.: ApJS, 116, 1 (1998) P. G. van Dokkum, M. Franx, D. Fabricant, et al.: ApJL, 520, L95 (1999) A. Vazdekis, A. J. Cenarro, J. Gorgas, et al.: MNRAS, 340, 1317 (2003) A. Vazdekis, H. Kuntschner, R. L. Davies, et al.: ApJL, 551, L127 (2001) Session IV Galaxies and cosmology Galaxy Evolution in Galaxy Clusters: Diffuse Light in the Virgo Cluster J. Alfonso L. Aguerri Instituto de Astrofı́sica de Canarias. C/ Vı́a Láctea s/n, 38200 La Laguna, Spain, jalfonso@iac.es Summary. I have reviewed, in the first part of this contribution, the different physical mechanisms driven the galaxy evolution in high density environment. The second part is focused on the discussion of the observational properties of the diffuse light in the Virgo cluster. I have also showed the main observational results of these cluster component obtained during the last years. 1 Introduction The visible universe is highly inhomogeneous at small scales, being the luminous matter condensed in stars, which are grouped forming galaxies. The galaxies can evolve in isolation, forming the so called field population, or they can be gravitationally bounded with other galaxies forming galaxy associations which goes from small groups up to massive galaxy clusters or superclusters. These are the biggest virialized structures known in the Universe. During decades, galaxy clusters have been used in order to determine largescale properties of the Universe, and confirm or refuse cosmological theories. They have also been used in order to understand the evolutive processes of galaxies; While field galaxies evolve passively, the evolution of galaxies in clusters strongly depends on the environment. One of the challenges of modern Astrophysics is to obtain a good theory about galaxy evolution, and one of the keys of this theory will be to explain the role placed by the environment. It has been known since the earliest observations or rich clusters that the properties of galaxies in clusters are quite distinct from field galaxies. The cluster population is dominated by early-type morphologies, primary ellipticals and S0s. They are the most abundant and homogeneous family of galaxies in clusters, following tight relations as: fundamental plane or the colour-magnitude relation. This type of galaxies dominate the cores of the clusters and should been in place before cluster virialization [16]. They have an old stellar population, which indicates that the population of Elliptical galaxies in clusters are evolving passively from high redshift. In contrast, late- 132 J. Alfonso L. Aguerri type galaxies in clusters are less abundant than in field and are located in the outermost regions where the local density is not very high. This distribution of morphologies is cluster is called the morphology density or morphologyclustercentric radius relations [17]. The differences between galaxies in field and clusters suggested that galaxies in clusters may have an intrinsically different formation process to field ones ([48, 26]). However, since the general acceptance of the hierarchical process as the preferred model of structure formation [32], in which bright galaxies would be the result of successive mergers and interactions, much attention has been focused on mechanisms that could transform late-type star forming galaxies in dense environments. There are several physical mechanisms, not present in the field, which can dramatically transform galaxies in high density environments. Galaxies in clusters can evolve due to, e.g. dynamical friction, which can slow down the more massive galaxies, circularise their orbits and enhance the merger rate ([30, 34]). Interactions with other galaxies and with the cluster gravitational potential can disrupt the outermost regions of the galaxies and produce galaxy morphological transformations from lateto early-types [38] or even change massive galaxies into dwarf ones [35]. Swept of cold gas produced by ram pressure stripping ([28, 44]) or swept of the hot gas reservoirs [10] can alter the star formation rate of galaxies in clusters. These mechanisms have been invoked in order to explain the differences between the photometrical components of cluster and field galaxies. Thus, it has been observed that the scale-lengths of the discs of spiral galaxies in the Coma cluster are smaller than those of similar galaxies in the field ([29, 2]). Interactions between galaxies and with the gravitational potential can disrupt the disks of spiral galaxies in clusters. They can be strong enough for transforming bright late-type spiral galaxies in dwarfs [4]. The disrupted material would be part of the intracluster light already detected in some nearby galaxy clusters ([6, 8, 3]) and galaxy groups ([14, 5]). 1.1 Diffuse light in galaxy clusters The study of the intracluster light (ICL) began with Zwicky’s ([55]) claimed discovery of an excess of light between galaxies in the Coma cluster. Its low surface brightness (≈ 28 mag arcsec−2 ) makes difficult to study the ICL systematically ([42, 52, 11, 27, 25]). Several approaches have been made during the last decades in order to study this elusive cluster component. We can mention the work done by [54], who stacked a large number of SDSS images to reach deep surface brightness levels. They identified the ICL as the excess of light at large galactrocentric distances over the r1/4 surface brightness profiles of cD galaxies. Other works identified the diffuse component as the residual light after the subtraction of the galaxies presented in deep surface photometry ([18, 20]). The central region of the Virgo cluster have been imaged by [37], observing that the ICL is located around the halos of the brightest galaxies in the Virgo cluster (see Fig 1). They also observed that ICL is formed by tidal Galaxy Evolution in Galaxy Clusters: Diffuse Light in the Virgo Cluster 133 tails, consequence of the destruction of galaxies in the centre of the cluster. Tidal tails around galaxies have also been observed in other galaxy clusters as the Coma cluster [1]. Another different approach for detecting the ICL consists on the direct detection of individuals stars in the intracluster region of nearby galaxy clusters. Three different kind of stars have been detected so far: SNe, RGB stars and PN. The first SNe detected in the intracluster region of the Virgo cluster between M86 and M84 was observed by [49]. In other clusters (Abell 403 and Abell 2122) two more SNe were detected by [24]. Recently, [41] found 6 SNe in the Fornax cluster, concluding that 16% of the light of Fornax is in the intracluster medium. Using high spatial resolve images from HST, [23] and [19] found a population of old evolved stars (RGB) in the Virgo cluster. This intracluster stellar population show a low metalicity and an old age of more than 10 Gyr [53]. They proposed that this metal poor stellar population come from disrupted dwarf galaxies of from the external regions of discs of spiral galaxies. All these works concluded that the ICL represent 15% of the light in the Virgo cluster. Another different approach was done by [9], who detected 3 PN not bounded to any galaxy in the Virgo cluster. They were free flying in the Virgo cluster gravitational potential. This opened a new approach to the study of the ICL in nearby galaxy clusters. Direct detection of intracluster PN (ICPN) have been observed in Virgo ([21, 22, 6, 3]) and Fornax clusters [51]. These works provide a consistent estimate that 10-20% of the light in galay clusters is located in the intracluster region. Fig. 1. Diffuse light in the Virgo cluster core. North is up; east is to the left. The white levels saturate at µV = 26.5, while the faintest features visible have a surface brightness of µV = 28.5. Figure taken from [37]. 134 J. Alfonso L. Aguerri Early theoretical studies predicted that the amount of the ICL should be a function of the galaxy number density [46]. These studies were based on analytic estimates of tidal stripping or simulations of individual galaxies orbiting in a smooth gravitational potential. Nowadays, cosmological simulations allow us to study in detail the evolution of galaxies in clusters environments. Thus, numerical simulations of galaxy cluster formation in ΛCDM cosmology show that the origin of the diffuse light is related with the formation of bright elliptical galaxies by mergers and the disruption of dwarf and bright galaxies by the cluster environment. The fraction of the ICL depends on the gravitational matter and the state of evolution of the cluster. Thus, the fraction of ICL was found to increase from 10-20% in clusters with 1014 M to up to 50% for very massive clusters [39]. For a fixed mass (1014 M ), [50] and [47] showed that the fraction of ICL increases also with the degree of dynamical evolution of the clusters. Those clusters with large differences in the magnitudes of the two brightest galaxies are evolved clusters and show the largest fraction of the diffuse light. [40] investigated the ICL for a Virgo-like cluster in one of these hierarchical simulations, predicting that the ICL is such clusters should be unrelaxed in velocity space and show significant substructures. The first radial velocity measurements for a substantial sample of ICPN [8] have indeed shown significant field-to-field variations and substructures. Our group is involved in a project for the detection of the intracluster light by the direct detection of ICPN in the Virgo cluster. We wish to detect the ICL at different radial distances from M87, being this project the first work of a extensive search of ICL in Virgo. In the following sections I will show the main results of the project during the last years. 2 Methodology The study of the ICL by the direct detection of ICPN has the advantage that these type of objects are easy to detect even at large distances. This is due to their strong emission in the [OIII]λ5007Å line. This has the advantages that detection of ICPN is possible with deep narrow-band images and that the ICPN radial velocities can be measured to investigate the dynamics of the ICL component. We have imaged 10 wide fields at different positions in the Virgo cluster (see Fig 2 to see the locations of some of the fields). These observations were taken with three wide field cameras: the Wide Field Imager on the ESO/MPI 2.2 m telescope, the Wide Field Camera at the 2.5 m Isaac Newton Telescope at the Roque the los Muchachos Observatory, and the Suprime-Cam at the prime focus of the Subaru 8.2 m Telescope. The images were acquired through a narrow band filter contained the wavelength of the [OIII] λ5007Å emission at the Virgo cluster mean redshift. In addition to this “on-band” filter, we also imaged in one broad-band filter (the “off-band” filter). For three of the fields we also have images taken with a narrow-band filter corresponding to the Hα emission at the redshift of the Virgo cluster. Galaxy Evolution in Galaxy Clusters: Diffuse Light in the Virgo Cluster 135 This allows us to detect the two strongest emission lines of PN (see [7] for more details of this technique). Fig. 2. Virgo Cluster core region with the positions of some of the fields studied. Figure taken from [3]. We have developed an automatic procedure for performing the photometry and identification of emission-line objects in our mosaic images in a homogeneous fashion (see [6, 3]). This procedure was applied to all fields in order to obtain the ICPN photometric candidates. The automatic extraction procedure stars with measuring the photometry of all objects in the images using SExtractor [12]. All objects are plotted in a colour-magnitude diagram (CMD), mn − mb versus mn , and are classified according to their positions in this diagram. The most reliable ICPN photometric candidates are point-like sources with no detected continuum emission and observed EW greater than 100 Å, after convolution with the photometric errors as a function of magnitude. Figure 3 shows the CMD for one of the observed fields. 2.1 Contamination of the catalogues It is possible that the ICPN samples may be contaminated by misclassified faint continuum objects, because of the selection based on a threshold in their 136 J. Alfonso L. Aguerri Fig. 3. Colour-magnitude diagram for all the sources in the RCN1 field. the horizontal line at mn − mv = −1 indicates objects with observed EW=110 Å. The diagonal line shows the magnitude corresponding to 1.0×σ above the sky in the V band. Full curved lines represent the 99% and 99.9% lines for the distribution of modelled continuum objects. The dashed lines represent 84% and 97.5% lines for the distribution of modeled objects with mn − mv = −1. The points are all objects detected by SExtractor. Diamonds are objects with a significant colour excess in the narrow-band filter. Figure taken from [6]. [OIII] fluxes. Due to the photometric errors in their [OIII] fluxes, some objects are assigned a flux brighter than their real flux. Because their LF also rises toward faint magnitudes, a significant number of objects will have measured [OIII] magnitudes brighter than the limiting magnitude. In addition, if their fluxes in the off-band image are bellow the limiting magnitude of that image, these objects will appear in the region of the CMD populated by the selected ICPN candidates, and they will be counted as ICPN candidates even though they are continuum objects. We call this “the spillover effect” from faint stars. Near to the limiting magnitude, the number of spillover stars will be negligible if an only if the off-band images is deep enough for detecting the weak continuum flux of these faint objects. This requires that the off-band image has and AB limiting magnitude of at least: mlim,b ≈ mlim,n + 3 < rms >, being < rms > the mean photometric error of objects with magnitude equal to the limiting magnitude of the narrow-band image (mlim,n ), amd mlim,b the limiting magnitude of the broad-band image. We have take into account this effect in our catalogues. Most of the broad-band images were deep enough, and the spillover effect was negligible. Galaxy Evolution in Galaxy Clusters: Diffuse Light in the Virgo Cluster 137 The photometric samples of ICPN can also be contaminated by emissionline background galaxies. This is because for [OII] starbusts galaxies at z=0.35 and Lyα galaxies at z≈ 3 their strong emission lines fall into our narrow-band filter width. As pointed out before, the threshold in EW implied by our selection criteria ensures that the selected ICPN photometric candidates are nearly free of [OII] emitters. However, some Lyα galaxies can contaminate our ICPN photometric samples. Spectroscopic follow-up observations of ICPN indeed found that a fraction of ICPN candidates were Lyα objects ([33, 8]). We have take into account this effect by the comparison of the LF of our ICPN photometrical candidates and those from emission line objects obtained using the same selection criteria as for the ICPN candidates which turned to be Lyα galaxies (see [14]). This comparison give us the number of background galaxies presented in our emission line catalogues. The number of contaminants strongly vary from field-to-field from a few to 100 per cent. 2.2 Spectroscopic follow-up Radial velocities of 40 ICPN in three fields (FCJ, Core and SUB) in the Virgo cluster core region were obtained with the multifiber FLAMES spectrograph on UT2 at the Very Large Telescope. These observations allowed us to confirm that a large fraction of the photometrically selected ICPN turned to be ICPN in the Virgo cluster. These data gave us the radial velocity distributions of ICPN and we could investigate the dynamical state of the Virgo cluster core region (see Fig 4). We have obtained that in the FCJ field, the velocity distribution of the PN is not consistent with a single Gaussian. It is dominated by a narrow peak, with vp = 1276 km s−1 and σp = 247 km s−1 , which we identify with the halo of M87 at a distance of ≈ 65 Mpc from the centre of M87. This proves the large extension of the halos of the brightest cluster galaxies. In the Core field, the distribution of ICPN line-of-sight velocities is clearly broader than the FCJ field and consistent with a Gaussian (see Fig 4). The Core field is in a region of Virgo devoid of bright galaxies but contains seven dwarf and three low-luminosity E/S galaxies near its Southwest borders. None of the confirmed ICPN lie within a circle of 3 times half the major-axis diameter of any of these galaxies, and there are no correlations of their velocities with the velocities of the nearest galaxies where these are known. Thus, in this field there is a clear intracluster stellar component (see Fig 4). In the SUB field, the velocity distribution from FLAMES spectra is again different from Core and FCJ. The SUB histogram could of LOS velocities shows substructures that are highly correlated with the systemic velocities of M86, M84 and NGC 4388. Most likely, all these PN belong to a very extended envelope around M84. It is possible that the somewhat low velocity with respect to M84 may be a sign of tidal stripping by M86 or of a recent merger with a smaller galaxies (see Fig 4). 138 J. Alfonso L. Aguerri Fig. 4. ICPN radial velocity distribution in the FCJ, Core and SUB pointings. In the FCJ panel, the dashed line shows a Gaussian with vrad = 1276 km s−1 and σrad = 247 km s−1 . In the Core panel, the dashed line shows a Gaussian with vrad = 1436 km s−1 and σrad = 538 km s−1 . In the SUB panel, the overploted dashed histogram shows the radial velocities from TNG spectroscopic follow-up [7]. The dashed red line shows a Gaussian with vrad = 1079 km s−1 and σrad = 286 km s−1 for the M84 peak. See for more details [8]. 3 Surface brightness and fraction of ICL in the Virgo cluster Determining the amount of the ICL from the observed numbers of ICPN is straightforward. From [36], if is the specific PN formation rate in PNs yr−1 L−1 , LT is the total bolometric luminosity of a sampled population, and tP N is the lifetime of a PN, which we take as 25000 yr, then the corresponding number of PNs, nP N , is: nP N = LT tP N . Theories of stellar evolution predict that the specific PN formation rate should be ≈ 2 × 10−11 stars yr−1 L−1 , nearly independent of population age or initial mass function [45]. Every stellar system should then have nP N = αLT = 50×10−8 PNs L−1 × LT . If the PN LF is valid 8 mag down to the cutoff, one can determine the fraction of PNs within 2.5 mag of the cutoff (M ∗ ) and thus define α2.5 as the number of PNs within 2.5 mag of M ∗ associated with a stellar population of total luminosity LT . Approximately one out of 10 of these PNs are within 2.5 mag of M∗ , and Galaxy Evolution in Galaxy Clusters: Diffuse Light in the Virgo Cluster 139 following from the above assumptions, most stellar populations should have α2.5 = 50 × 10−9 PNs L−1 [22]. Thus, the observed number of ICPNs can be used to infer the total luminosity of the parent stellar population. Fig. 5. Number density of PNs (top) and surface brightness of the ICL (bottom) in our surveyed fields. In both panels, circles and diamonds represent our Virgo fields, triangles show the measurements of the ICL from RGB stars [19]. The asterisks and the full line show the B-band luminosity of Virgo galaxies averaged in rings [13]. Distances are relative to M87. However, as first noticed by [43], observations of PN samples in galaxies show that α2.5 varies strongly as a function of colour. [31] found that α2.5 decreases by a factor of 7 from the value 50×10−9 PN L−1 measured in dwarf elliptical galaxies like NGC 205 to 7×10−9 PN L−1 observed in Virgo elliptical galaxies. The amount of the ICL depends directly on the adopted value of α, which is thus not very well constrained and is a function of the (B-V) colour of the parent stellar population, currently unknown for the Virgo ICL. To take this uncertainty into account in our estimates of the intracluster luminosity in the different fields, we consider three plausible values for α, which are (1) the value appropriate for and evolved population like that of the M31 bulge [15], (2) the value determined by [19] for the intracluster red giant branch stars 140 J. Alfonso L. Aguerri observed with HST, and (3) the value determined from the [31] empirical relation. In this case the B-V of the parent stellar population is determined from the average colours of the Virgo galaxies near the field position. Figure 5 shows the resulting values of the B-band surface brightness and PN number density as a function of the distance from M87 for our Virgo cluster fields (full points and diamonds). It is also overploted the surface brightness of the ICL from the detection of RGB stars (triangles; [19]), and the surface brightness of the galaxies in the cluster (asterisks; [13]). Notice that the ICL is concentrated within a radius of 120 arcmin from M87. The emission line objects detected out from this radius are compatible with being background galaxies. In the central region of the cluster the ICL represent up to ≈ 10% of the total light in the cluster [3]. This indicates that the diffuse light could be associated with the formation of the large galaxy halos of bright elliptical galaxies presented in the central region of the Virgo cluster. Figure 5 also shows that there is a large field to field variation. This can indicate that the ICL is still not relaxed in the cluster potential. There has not been enough time for phase mixing to erase these variations. This places a strong constraint on the age and origin of the ICL in the Virgo cluster, which must have been brought into this location not much more than a few dynamical times ago. This result indicates that the physical mechanism responsible of the formation of the diffuse component in the Virgo cluster is active until recent epochs of the cluster evolution, discarding a primordial origin of the ICL. Unfortunately, there are no other galaxy cluster as rich as Virgo at small distances from the Milky Way. The technique used in Virgo for the detection of ICPNs can not be applied to other galaxy clusters at larger distances of more than ≈ 25M pc.. The new generation of telescopes, as GTC, will give us the opportunity to observe the diffuse light in other cluster but with different techniques. Blind spectroscopic surveys of emission line objects will be possible in 10 m class telescopes for galaxy clusters located closer than ≈ 100 Mpc. This will give us the opportunity to study the properties of the diffuse component in other clusters with different physical conditions, and test the theoretical results about the dependence of the properties of the ICL and the dynamical state of the cluster. References 1. Adami, C., et al. A&A, 429, 39 (2005) 2. Aguerri, J. A. L., Iglesias-Paramo, J., Vilchez, J. M., & Muñoz-Tuñón, C. AJ, 127, 1344 (2004) 3. Aguerri, J. A. L., Gerhard, O. E., Arnaboldi, M., Napolitano, N. R., CastroRodriguez, N., & Freeman, K. C. AJ, 129, 2585 (2005) 4. Aguerri, J. A. L., Iglesias-Páramo, J., Vı́lchez, J. M., Muñoz-Tuñón, C., & Sánchez-Janssen, R. AJ, 130, 475 (2005) Galaxy Evolution in Galaxy Clusters: Diffuse Light in the Virgo Cluster 141 5. Aguerri, J. A. L., Castro-Rodrı́guez, N., Napolitano, N., Arnaboldi, M., & Gerhard, O. A&A, 457, 771 (2006) 6. Arnaboldi, M., et al. AJ, 123, 760 (2002) 7. Arnaboldi, M., et al. AJ, 125, 514 (2003) 8. Arnaboldi, M., Gerhard, O., Aguerri, J. A. L., Freeman, K. C., Napolitano, N. R., Okamura, S., & Yasuda, N. ApJ, 614, L33 (2004) 9. Arnaboldi, M., et al. ApJ, 472, 145 (1996) 10. Bekki, K., Couch, W. J., & Shioya, Y. ApJ, 577, 651 (2002) 11. Bernstein, G. M., Nichol, R. C., Tyson, J. A., Ulmer, M. P., & Wittman, D. AJ, 110, 1507 (1995) 12. Bertin, E., & Arnouts, S.A&AS, 117, 393 (1996) 13. Binggeli, B., Tammann, G. A., & Sandage, A. AJ, 94, 251 (1987) 14. Castro-Rodrı́guez, N., Aguerri, J. A. L., Arnaboldi, M., Gerhard, O., Freeman, K. C., Napolitano, N. R., & Capaccioli, M. A&A, 405, 803 (2003) 15. Ciardullo, R., Jacoby, G. H., Ford, H. C., & Neill, J. D. ApJ, 339, 53 (1989) 16. Dressler, A., et al. ApJ 490, 577 (1997) 17. Dressler, A. ApJS, 42, 565 (1980) 18. Feldmeier, J. J., 19. Durrell, P. R., Ciardullo, R., Feldmeier, J. J., Jacoby, G. H., & Sigurdsson, S. ApJ, 570, 119 (2002) Mihos, J. C., Morrison, H. L., Rodney, S. A., & Harding, P. ApJ, 575, 779 (2002) 20. Feldmeier, J. J., Mihos, J. C., Morrison, H. L., Harding, P., Kaib, N., & Dubinski, J. ApJ, 609, 617 (2004) 21. Feldmeier, J. J., Ciardullo, R., Jacoby, G. H., & Durrell, P. R. ApJS, 145, 65 (2003) 22. Feldmeier, J. J., Ciardullo, R., Jacoby, G. H., & Durrell, P. R. ApJ, 615, 196 (2004) 23. Ferguson, H. C., Tanvir, N. R., & von Hippel, T. Nat, 391, 461 (1998) 24. Gal-Yam, A., Maoz, D., Guhathakurta, P., & Filippenko, A. V. AJ, 125, 1087 (2003) 25. Gonzalez, A. H., Zabludoff, A. I., Zaritsky, D., & Dalcanton, J. J. ApJ, 536, 561 (2000) 26. Gott, J. R., III, & Thuan, T. X. ApJ, 204, 649 (1976) 27. Gregg, M. D., & West, M. J. Nat, 396, 549 (1998) 28. Gunn, J. E., & Gott, J. R. I. ApJ, 176, 1 (1972) 29. Gutiérrez, C. M., Trujillo, I., Aguerri, J. A. L., Graham, A. W., & Caon, N. ApJ, 602, 664 (2004) 30. den Hartog, R., & Katgert, P. MNRAS, 279, 349 (1996) 31. Hui, X., Ford, H. C., Ciardullo, R., & Jacoby, G. H. ApJ, 414, 463 (1993) 32. Kauffmann, G., & White, S. D. M. MNRAS, 261, 921 (1993) 33. Kudritzki, R.-P., et al. ApJ, 536, 19 (2000) 34. Mamon, G. A. ApJ, 401, L3 (1992) 35. Mastropietro, C., Moore, B., Mayer, L., Debattista, V. P., Piffaretti, R., & Stadel, J. MNRAS, 364, 607 (2005) 36. Mendez, R. H., Kudritzki, R. P., Ciardullo, R., & Jacoby, G. H. A&A, 275, 534 (1993) 37. Mihos, J. C., Harding, P., Feldmeier, J., & Morrison, H. ApJ, 631, L41 (2005) 38. Moore, B., Katz, N., Lake, G., Dressler, A., & Oemler, A., Jr., Nat , 379, 613 (1996) 142 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. 55. J. Alfonso L. Aguerri Murante, G., et al. ApJ, 607, L83 (2004) Napolitano, N. R., et al. ApJ, 594, 172 (2003) Neill, J. D., Shara, M. M., & Oegerle, W. R. ApJ, 618, 692 (2005) Oemler, A. ApJ, 180, 11 (1973) Peimbert, M. Revista Mexicana de Astronomia y Astrofisica, 20, 119 (1990) Quilis, V., Moore, B., & Bower, R. Science, 288, 1617 (2000) Renzini, A & Buzzoni, A.. In: Spectral energy Distributions of Galaxies, ed by C. Chiosy & A. Renzini(Dordrecht: Reidel) pp 195 (1986) Richstone, D. O., & Malumuth, E. M. ApJ, 268, 30 (1983) Rudick, C. S., Mihos, J. C., & McBride, C. ApJ, 648, 936 (2006) Sandage, A., Freeman, K. C., & Stokes, N. R. ApJ, 160, 831 (1970) Smith, H. A. AJ, 86, 998 (1981) Sommer-Larsen, J. MNRAS, 369, 958 (2006) Theuns, T., & Warren, S. J. MNRAS, 284, L11 (1997) Thuan, T. X., & Kormendy, J. PASP, 89, 466 (1977) Williams et al. ApJ, in press (2006) Zibetti, S., White, S. D. M., Schneider, D. P., & Brinkmann, J.: MNRAS, 358, 949 (2005) Zwicky, F. PASP, 63, 61 (1951) The quest for obscured AGN at cosmological distances: Infrared Power-Law Galaxies A. Alonso-Herrero1,2, J.L. Donley2 , G.H. Rieke2 , J.R. Rigby3,2 and P.G. Pérez-González4,2 1 2 3 4 DAMIR, Instituto de Estructura de la Materia, CSIC, 28006 Madrid, Spain, aalonso@damir.iem.csic.es Steward Observatory, University of Arizona, Tucson, AZ 85721, USA Carnegie Observatories, Pasadena, CA 91101, USA Departamento de Astrofı́sica y CC de la Atmósfera, UCM, 28040 Madrid, Spain Summary. We summarize multiwavelength properties of a sample of galaxies in the Chandra Deep Field North (CDF-N) and South (CDF-S) whose Spectral Energy Distributions (SEDs) exhibit the characteristic power-law behavior expected for AGN in the Spitzer/IRAC 3.6−8 µm bands. AGN selected this way tend to comprise the majority of high X-ray luminosity AGN, whereas AGN selected via other IRAC color-color criteria might contain more star-formation dominated galaxies. Approximately half of these IR power-law galaxies in the CDF-S are detected in deep (1 Ms) Chandra X-ray imaging, although in the CDF-N (2 Ms) about 77% are detected at the 3 σ level. The SEDs and X-ray upper limits of the sources not detected in X-rays are consistent with those of obscured AGN, and are significantly different from those of massive star-forming galaxies. About 40% of IR power-law galaxies detected in X-rays have SEDs resembling that of an optical QSO and morphologies dominated by bright point source emission. The remaining 60% have SEDs whose UV and optical continuum are much steeper (obscured) and more extended morphologies than those detected in X-rays. Most of the IR power-law galaxies not detected in X-rays have IR (8 − 1000 µm) above 1012 L , and X-ray (upper limits) to mid-IR ratios similar to those of local warm (ie, hosting an AGN) ULIRGs. The SED shapes of power-law galaxies are consistent with the obscured fraction (4:1) as derived from the X-ray column densities, if we assume that all the sources not detected in X-rays are heavily absorbed. IR power-law galaxies may account for between 20% and 50% of the predicted number density of mid-IR detected obscured AGN. The remaining obscured AGN probably have rest-frame SEDs dominated by stellar emission. 1 Introduction Active Galactic Nuclei (AGN) are sources of luminous X-ray emission, and at cosmological distances AGN are routinely selected from deep X-ray (< 10 keV) exposures ([8]). Highly obscured (NH > 1023 − 1024 cm−2 ) AGN are thought to be a major contributor to the hard X-ray background ([24, 9, 39, 144 Alonso-Herrero et al. 40]). However, the majority of them might not be detected in these X-ray surveys because a large fraction of their soft X-ray, UV, and optical emission is absorbed, and presumably reradiated in the infrared (IR). Numerous attempts have been made to detect this population of heavily obscured AGN, many of which have focused on the MIR emission where the obscured radiation is expected to be reemitted (e.g., [4, 20, 32, 28, 11]), or on combinations of MIR and multiwavelength data (e.g., [10, 23]). In the mid-IR, AGN can often be distinguished by their characteristic power-law emission (e.g.,[25, 13]). This emission is not necessarily due to a single source, but can arise from the combination of non-thermal nuclear emission and thermal emission from various nuclear dust components ([29]). We summarize here the properties of galaxies showing the characteristic power-law behavior expected for AGN in the Spitzer 3.6 − 8 µm bands detected in the Chandra Deep Field North and South (CDF-N and CDF-N) studied in [4] and [11], respectively. We also discuss results from other IR-based methods to detect high-z obscured AGN. We use H0 = 71 km s−1 Mpc−1 , ΩM = 0.3, and ΩΛ = 0.7. 2 The Sample of IR power-law galaxies 2.1 Selection We chose two cosmological fields with deep X-ray coverage (CDF-N: 2 Ms and CDF-S: 1 Ms, see [1] for details) to look for obscured AGN. We selected as power-law galaxies sources that were detected in each of the four IRAC (3.6, 4.5, 5.8, and 8 µm) bands and whose IRAC spectra could be fitted as fν ∝ ν α , where α is the spectral index. We used a minimum χ2 criterion to select galaxies whose IRAC SEDs followed a power law with spectral index α < −0.5. The choice for the spectral index was based on the empirical spectral energy distributions (SEDs) of bright QSO selected in the optical, X-rays and near-IR (e.g., [25, 13, 18, 19]) and Seyfert galaxies (e.g., [36, 12]). There are two slight differences in the two catalogs of IR power-law galaxies. In the CDF-S [4] started the selection of the power-law candidates from the 24 µm catalog of [26], without any further requirements on the flux limits of the IRAC catalogs. [11] in the CDF-N instead imposed a strict S/N= 6 flux density cut in each of the IRAC bands but did not require a 24 µm detection, although virtually all of the power-law galaxies in the CDF-N were also detected at 24 µm down to ∼ 80 µJy (equivalent to the 80% completeness limit of the CDF-S, see [26]). To minimize the chances of selecting non-active galaxies we constructed optical-MIR SEDs (see §4), and compared them with theoretical and observational templates of star-forming galaxies. We rejected any source selected via the power-law criteria whose SED resembled a star-forming galaxy. The final samples included 92 and 62 galaxies in the CDF-S and CDF-N, respectively (see [4] and [11] for details). Infrared Power-Law Galaxies 145 Fig. 1. Location of the CDF-N IR power-law galaxies (diamonds) on the Lacy et al. (left) and Stern et al. (right) IRAC color-color diagrams together with X-ray sources (crosses from [1]) and IRAC galaxies (dots, non power-law galaxies) detected in this field. The shaded regions indicate the AGN loci, and the straight lines within them the power-law criterion. We also show the z-evolution (z = 0 − 2.5 where z = 0 is indicated by the large symbol at the edge of each template line) of different templates. The line marked with upside down triangle corresponds to a starburst ULIRG, triangle Arp 220, square an AGN ULIRG, filled dot the average of the radio-quiet QSOs of [13], and star symbol a star-forming galaxy. A small (∼ 25 − 30%, depending on the field) fraction of the power-law galaxies, typically the optically bright X-ray sources (see next section), have spectroscopic redshifts (e.g., [33] in CDF-S). We supplemented the available spectroscopic redshifts with photometric ones estimated with an improved version of the method described by [27]. We find that the IR power-law galaxies tend to lie at significantly higher redshifts (z > 1) than the X-ray sources (median z ∼ 0.7, see [8] for a review) in both fields. 2.2 Comparison between IRAC power-law and color-color criteria [20] and [32] defined AGN selection criteria based on Spitzer/IRAC color-color diagrams. The Lacy et al criterion is based on SDSS QSOs, and therefore excludes AGN in which the host galaxy dominates the MIR (see e.g., [3, 31, 14, 28]), as well as AGN obscured in the mid-IR. Our power-law galaxies fall along a straight line well within the Lacy et al diagram, although they do not cover completely the available color space (see Fig. 1). The Stern et al criteria are based on the observed properties of spectroscopically classified AGN, and provide a closer match to our power-law technique. While the colorcolor selected samples comprise a higher fraction of the low X-ray luminosity AGN than does the power-law selected sample, the color criteria select more 146 Alonso-Herrero et al. Fig. 2. Left panel: COMBO-17 ([38]) R-band versus the Chandra full band flux of IR power-law galaxies in the CDF-S (similar results are found for the CDF-N). When available we show the optical spectroscopic classifications from [33] (see §5 for details). The galaxies in our sample not detected by COMBO-17 are shown as upper limits at R = 27 mag. Right panel: Rest-frame 2 − 8 keV luminosity vs hard X-ray to 12 µm for IR power-law galaxies (asterisks and X-ray upper limits; the circles denote those classified as BLAGN) in the CDF-S compared with local cool (no AGN) and warm (hosting an AGN) ULIRGs. sources not detected in X-rays, due at least in part to a higher degree of contamination from ULIRGs dominated by star formation (Fig. 1 and [6]). Recently [21] obtained follow-up optical spectroscopy of objects selected according to [20] in the Spitzer First Look Survey and SWIRE XMM-LSS fields. Their sample is flux-limited at 24 µm, although their objects are much brighter (f24µm = 4 − 20 mJy, median of 5 mJy, and median R-band magnitudes R ∼ 18 mag) than our power-law galaxies (f24µm ∼ 0.08 − 3 mJy, and R ∼ 23 mag for those detected by COMBO-17 see next section), and on average their AGN are closer zsp ∼ 0.6 compared with z ∼ 1.5 for our power-law galaxies. The location of this sample on the [20] IRAC color-color diagram (figure 7 in [21]) is almost identical to the positions of the power-law galaxies shown in Fig 1 (right panel). Their selection technique has proven to be very effective at selecting AGN as their follow-up spectroscopy shows that approximately 90% have AGN signatures with one-third of them showing broad-line regions, thus an obscured-to-unobscured ratio of 2:1 (see §5 for the power-law galaxies). All these properties seem to indicate that these color-color selected galaxies represent the brightest end of the power-law galaxies. Infrared Power-Law Galaxies 147 Fig. 3. Rest-frame SEDs (filled circles) of CDF-S IR power-law galaxies detected in X-rays with spectroscopic redshifts and classifications from [33] (IDs given on the right-hand side of each panel). For each SED class (only the BLAGN and NLAGN classes are shown, see [4] for more details) we have constructed an average template, shown as the solid line in each panel. 3 X-ray, Infrared, and Optical Properties In both CDF-N and CDF-S we found that approximately 50% of the IR power-law galaxies were detected in at least one Chandra band using the X-ray catalogs of [1]. In the CDF-S we stacked the X-ray data of a few individually undetected galaxies (at off-axis angles of θ < 7.50 ) and found a significant detection in the hard-band (3.1 σ) and a tentative detection in the soft band (2 σ). For z = 2 these would correspond to observed soft and hard luminosities of < 7 × 1041 erg s−1 and 4 × 1042 erg s−1 , respectively. This is consistent with obscured AGN. Since the X-ray exposure of the CDF-N is twice that of CDF-S we searched for faint X-ray emission at the positions of the power-law galaxies not in the [1] catalog. We found that the X-ray detection rate increases to 77% at the 3 σ level. The power-law galaxies make up a significant fraction of the high X-ray luminosity sample, as our selection criteria require the AGN to be energetically dominant. The lower luminosity X-ray sources not identified as power-law galaxies tend to be dominated by the 1.6 µm stellar bump in the optical to near-IR bands (see also [3, 14, 28, 31]). A large fraction of IR luminous high-z galaxies have been found to host AGN (e.g., SCUBA galaxies [2]), and IR luminous galaxies at z ∼ 1 − 2 (e.g., [42]). We measured the total IR (8−1000 µm) luminosity of the CDF-S powerlaw galaxies from the rest-frame 12 µm luminosity. Although our procedure to compute IR luminosities is similar to that of [27], we took special care to use 12 µm to IR luminosity ratios specific to the class of galaxies in study. In particular, galaxies whose SEDs resemble those of optical QSOs (see §4 and Fig. 3) show 12 µm to IR luminosity ratios significantly lower than the typical values of cool ULIRGs and some warm ULIRGs (e.g., Mrk 231). All the IR power-law galaxies are highly luminous. About 30% are in the hyperluminous 148 Alonso-Herrero et al. Fig. 4. Examples of the observed SEDs (filled circles and star symbols) of CDF-S IR power-law galaxies not detected in X-rays. Each galaxy is shown with the closest average template constructed using the X-ray detected ones (see [4] for details). class (LIR > 1013 L ), 41% are ULIRGs (LIR = 1012 − 1013 L ), and all but one of the rest are LIRGs (LIR = 1011 − 1012 L ). At the lower IR luminosity end (LIR < 1012 L ) a large fraction are detected in X-rays and tend to have SEDs similar to those of optical QSOs (see next section). About one-quarter of the CDF-S IR power-law galaxies are optically faint i.e., were not detected by COMBO-17 ([38]) down to a limit of R ∼ 26.5 mag. Moreover, the fraction of power-law galaxies not detected in X-rays increases toward fainter R-band magnitudes (see Fig. 2), an indication of their obscured nature as also revealed by their SEDs (see also §5). A number of works (e.g., [7] and references therein) have demonstrated that X-ray to optical flux ratios can be useful for distinguishing between AGN and star-forming galaxies for sources detected in deep X-ray exposures. Fig. 2 (left panel) shows that the majority of the galaxies (or their upper limits) in the CDF-S are consistent with being AGN or transition objects based on the X-ray vs. R-band diagram (similar results are found for the CDF-N galaxies). The location of the IR power-law galaxies on this diagram (see [7]) indicates X-ray luminosities (or Infrared Power-Law Galaxies 149 upper limits) above 1041 erg s−1 , as also shown by the right panel of Fig. 2. The rest-frame hard X-ray/12 µm ratios of the IR power-law galaxies are similar to those of local warm ULIRGs (i.e., those containing an AGN) and QSO. Fig. 5. Examples of rest-frame SEDs of CDF-S VVDS galaxies at z > 1 included in the sample of predominantly star-forming galaxies selected at 24 µm by [27]. 4 SEDs and Morphologies Using available multiwavelength datasets (see [4, 11] for references) of the two cosmological fields we constructed SEDs for our sample. The spectroscopic classifications (broad vs. narrow lines1 ) tend to agree with two distinct types of SEDs. About 40% of the CDF-S IR power-law galaxies detected in X-rays are classified as BLAGN and have SEDs (Fig. 3) similar to the average radioquiet QSO SED of [13], that is, with an optical–to–mid-IR continuum almost flat in νfν with a UV bump. The remaining X-ray sources with narrow lines (NLAGN) have SEDs similar to the BLAGN but their UV and optical continua are much steeper (obscured), and some of them resemble local warm ULIRGs. The majority of the power-law galaxies not detected in X-rays (Fig. 4) have steep SEDs similar to the NLAGN or ULIRG class as they tend to be optically fainter (see §3) and possibly more obscured (see §5) than the X-ray sources. In contrast with our power-law galaxies, massive galaxies from the VIRMOS VLT Deep Survey (VVDS, [22]) at 1 < z < 2 with 24 µm detections from the sample of [27] are predominantly star-forming galaxies with a prominent stellar bump at 1.6 µm due to an evolved (red giants and supergiants) stellar population. Moreover, the SEDs of power-law galaxies are also significantly different from those of the majority of optically-dull AGN in the CDF-S which show SEDs dominated by stellar light originating in the host galaxy (see [31]). 1 The Szokoly et al. spectral classifications of CDF-S X-ray sources with clear AGN signatures were: BLAGN (broad-line AGN) and HEX (high excitation lines). Approximately 50% of their X-ray sources did not have a clear AGN signature in their optical spectra: LEX (low excitation lines, also termed optically-dull AGN and X-BONGS) and ABS (absorption lines). 150 Alonso-Herrero et al. Fig. 6. Postage stamps of optically-bright power-law galaxies in the CDF-S. The data are from the public release of the GOODS HST/ACS F606W images. Examples of X-ray detected power-law galaxies are shown in the top and middle panels with the IDs, and spectroscopic redshifts and classifications from [33]. Power-law galaxies not detected in X-rays are shown in the bottom panel. Fig. 4 shows the Great Observatories Origins Deep Survey (GOODS) ACS ([15]) observed optical morphologies of optically-bright power-law galaxies in the CDF-S. The power-law galaxies spectroscopically classified as BLAGN (and SED type) display bright nuclear point sources suggesting that the optical light is dominated by the AGN, which is consistent with the fact that our power-law criteria selects the most X-ray luminous AGN (see also [31]). The power-law galaxies without broad lines (those classified as HEX or LEX by [33]) have disky or irregular morphologies. A morphological characterization of the IR power-law galaxies not detected in X-rays is difficult as only about half of them are detected in the GOODS/ACS images and they are faint. As can be seen for a few examples in Fig. 4 they have irregular, knotty, and/or interacting morphologies, and do not appear to contain bright point sources. 5 Obscuration and Obscured Fraction In the distant universe the X-ray background and luminosity synthesis models predict global obscured (NH ≥ 1022 cm−2 ) to unobscured ratios of 3:1 to 4:1 (e.g., [9, 16]), significantly higher than the observed ratios of spectroscopically Infrared Power-Law Galaxies 151 Fig. 7. Left panel. Distribution of X-ray column densities for the CDF-N power-law galaxies that are cataloged X-ray sources (empty) and those only weakly detected in X-rays (shaded) from [11]. We also include those CDF-S power-law galaxies selected by [4] for which [31] measured a column density. Right Panel. NH distribution for all the CDF-S X-ray optically active AGN (empty) with spectroscopic classifications [33, 31], excluding all the power-law galaxies shown in the left panel. The shaded histogram shows the NH of the optically-dull AGN (only those classified as LEX). identified X-ray sources in deep fields (e.g., [5, 33]) including those detected in the mid-IR (e.g., [30]). [11] estimated the intrinsic column densities of each of the X-ray well detected and weakly X-ray emitting power-law galaxies in the CDF-N. We also included a few CDF-S IR power-law galaxies ([4]) for which [31] estimated the X-ray column densities. The column density distribution of the power-law galaxies (Fig. 7) is significantly different from that of optically bright X-ray AGN with spectroscopic classifications (see also [34]), including the optically-dull AGN which are believed to suffer strong obscuration ([31]). From Fig. 7 it is clear that the weakly detected IR power-law galaxies are consistent with being obscured (NH ∼ 1022 − 1024 cm−2 ) but not Comptonthick (NH ≥ 1024 cm−2 ). If all the X-ray non-detected power-law galaxies are obscured, the maximum obscured ratio is 4:1 (for the CDF-N power-law galaxies). This NH -based obscured fraction of power-law galaxies agrees well with the ratio of BLAGN (unobscured) SED vs. NLAGN (obscured) SEDs found in the CDF-S. We can finally estimate the ratio of obscured to unobscured mid-IR detected AGN in the CDF-S. The unobscured AGN are all those X-ray sources (detected in the hard band to make sure they are AGN) with NH < 1022 cm−2 , whereas in the obscured category we include all obscured X-ray sources with NH > 1022 cm−2 , and all the obscured IR power-law galaxies. We find an observed ratio of obscured to unobscured AGN of 2:1 in the CDF-S. Comparing with the predictions of [35] for the 24 µm detected AGN number density we 152 Alonso-Herrero et al. find that our sample of power-law galaxies only accounts for approximately 20% of all the mid-IR emitting obscured AGN in the CDF-N. This fraction can be as high as ∼ 50% in the CDF-S as a result of larger sample of powerlaw galaxies there (we did not impose IRAC high S/N detections, see §2 and [4, 11] for details). The remainder should have SEDs dominated by or strongly affected by the host galaxy or red power-law SEDs that fall below the IRAC detection limit. This is not surprising as our power-law criteria require the AGN to be energetically dominant in the near to mid-IR. 6 Other Infrared-Based Searches for Obscured AGN The selection of IR-bright optically-faint galaxies has been suggested as another method for identifying obscured AGN ([17, 37]), although the selection criteria (R > 23.9 and f24 µm > 0.75 − 1 mJy) were set so that IRS follow-up spectroscopy could be obtained. These criteria select mostly high-z (median z ∼ 2.2) galaxies, with only a small fraction showing the characteristic aromatic feature emission of star formating galaxies. The majority have IRS spectra similar to local AGN-dominated ULIRGs with either a featureless power-law rest-frame mid-IR continuum or deep silicate features at 9.7 µm. In addition, their SEDs lack a strong 1.6 µm stellar bump, and are similar to those of IR power-law galaxies. Their properties are consistent with being optically obscured AGN-powered ULIRGs with LIR > 1012 L (see [17, 37] for details). Only a few galaxies in the CDF-N and CDF-S fall within the [17, 37] flux density cuts. We can instead compare with the 24 µm/8 µm vs. 24 µm/R-band diagram criterion proposed by [41, 42] to select obscured AGN. The location of our CDF-N power-law galaxies on this diagram can be seen in Fig. 8. We also show the comparison X-ray sources (from [1]) and other IRAC sources in the field that do not meet the power-law criteria. We find that all these samples cover a large range in colors, but the power-law galaxies comprise a significant fraction (∼ 30 − 40%) of the highly optically reddened members of the comparison X-ray sample. This suggests that the power-law selection is capable of detecting both optically obscured and unobscured AGN, and that a large fraction of the IR-bright/optically-faint sources in the comparison sample have power-law SEDs in the near and mid-IR (Fig. 8). Radio emission is another good way to select AGN as it is unaffected by dust absorption. [10] in the CDF-N used a radio to mid-IR ratio to select galaxies that are too bright in radio to be star-forming galaxies. They found that ∼ 30% of their radio-loud AGN are not detected in X-rays suggesting strong obscuration. [23] looked for a population of radio intermediate and radio quiet AGN by selecting 24 µm sources (f24 µm ∼ 0.3 − 1 mJy) with radio emission and imposed a flux density cut at 3.6 µm to filter out type-1 and radio-loud QSOs. They found a population of QSOs at 1.4 < z < 4.2 (median z = 2, the epoch of QSO maximum activity) with a ratio of obscured to Infrared Power-Law Galaxies 153 Fig. 8. Location of the CDF-N power-law galaxies and comparison samples (symbols as in Fig. 1) on the color-color diagram of [41] (left panel) where the dotted lines indicate the Yan et al. selection criteria for possible dust-reddened AGN. The right panel shows the power-law fraction of the IRAC comparison sample (non power-law galaxies) as a function of the 24 µm over R-band flux ratio. unobscured of (2 − 3):1, and postulated that this population of obscured AGN may be responsible for most of the black hole growth in the young universe. 7 Summary Deep X-ray cosmological surveys are efficient at detecting AGN at high-z but they only account for a ∼ 1:1 ratio of obscured-to-unobscured AGN, whereas synthesis models of the X-ray background require ratios of between 3:1 and 4:1. Thus there is a significant population of obscured (NH > 1023 − 1024 cm−2 ) AGN being missed by current deep X-ray (< 10 keV) observations. We describe searches for this population. Our selection criterion is based on the characteristic IR power-law emission shown by local QSOs. By selecting galaxies with power-law emission in the Spitzer/IRAC bands (3.6, 4.5, 5.8, and 8 µm) we avoid high-z galaxies whose SEDs are dominated by stellar emission and star formation peaking at 1.6 µm. Only ∼ 50% of IR power-law galaxies are detected in deep (1−2 Ms) X-ray exposures. This fraction increases to 75% if we include weakly detected X-ray sources at the 3 σ level in the field with the deepest X-ray exposure (CDF-N). The optical (faint), IR (mostly ULIRGs and hyper luminous IR galaxies), and X-ray properties, the X-ray column densities NH (moderately obscured, but not Compton-thick) and redshift (z > 1) distributions, and SED shapes of a large fraction (up to 80%) of IR power-law galaxies are significantly different from bright X-ray selected AGN. This may indicate that a large fraction of IR power-law galaxies are good candidates to host obscured AGN, and could account for a ratio of 2:1 of obscured-to-unobscured AGN at high-z. Other mid-IR based criteria (e.g., [10, 17, 20, 21, 23, 32, 41, 42, 37]) are also finding populations of obscured AGN. There might be a significant overlap between populations of bright mid-IR AGN selected with all these methods, 154 Alonso-Herrero et al. and thus a complete census of the entire obscured AGN population is needed to determine whether we can account for the obscured fraction of the X-ray background. Acknowledgements: A. A. H. acknowledges support from the Spanish Plan Nacional del Espacio under grant ESP2005-01480 and P. G. P.-G. from the Spanish Programa Nacional de Astronomı́a y Astrofı́sica under grant AYA 2004-01676 and the Comunidad de Madrid ASTRID I+D project. Support for this work was also provided by NASA through Contract no. 960785 and 1256790 issued by JPL/Caltech. References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. Alexander D.M. et al: AJ 126, 539 (2003) D.M. Alexander D.M. et al: Nature 434, 738 (2005) Alonso-Herrero A. et al: ApJS 154, 155 (2004) Alonso-Herrero A. et al: ApJ 640, 167 (2006) Barger A. et al: AJ 126, 632 (2003) Barmby P. et al: ApJ 642, 126 (2006) Bauer F.E. et al: AJ 128, 2048 (2004) Brandt W.N. & G. Hasinger G.: ARA&A 43, 827 (2005) Comastri A. et al: A&A 296, 1 (1995) Donley J.L. et al: ApJ 634, 169 (2005) Donley J.L. et al: ApJ submitted (2006) Edelson R.A. et al: ApJ 321, 233 (1987) Elvis M. et al: ApJS 95, 1 (1994) Franceschini A. et al: AJ 129, 2074 (2005) Giavalisco M. et al: ApJ 600, L93 (2004) Gilli R.: Advances in Space Research 34, 2470 (2004) Houck J. et al: ApJ 622, L105 (2005) Ivezić Z. et al: AJ 124, 2364 (2002) Kuraszkiewicz J.K. et al: ApJ 590, 128 (2003) Lacy M. et al: ApJS 154, 166 (2004) Lacy M. et al: ApJ in press, astro-ph/0609594 (2006) Le Fèvre O. et al: A&A 428, 1043 (2004) Martı́nez-Sansigre A. et al: Nature 436, 666 (2005) Mushotzky R.F. et al: Nature 404, 459 (2000) Neugebauer G. et al: ApJ 230, 79 (1979) Papovich C. et al: ApJS 154, 70 (2004) Pérez-González P.G. et al: ApJ 630, 82 (2005) Polletta M.C. et al: ApJ 642, 673 (2006) Rieke G.H. & M.J. Lebofsky: ApJ 250, 87 (1981) Rigby J.R. et al: ApJS 154, 160 (2004) Rigby J.R. et al: ApJ 645, 115 (2006) Stern D. et al: ApJ 631, 136 (2005) Szokoly G.D. et al: ApJS 155, 271 (2004) Tozzi P. et al: A&A 451, 457 (2006) Treister E. et al: ApJ 616,123 (2004) Ward M.J. et al: ApJ 315, 74 (1987) Infrared Power-Law Galaxies 37. 38. 39. 40. 41. 42. Weedman D. et al: ApJ 651, 101 (2006) Wolf C. et al: A&A 421, 193 (2004) Worsley M.A. et al: MNRAS 352, L28 (2004) Worsley M.A. et al: MNRAS 357, 1281 (2005) Yan L. et al: ApJS 154, 60 (2004) Yan L. et al: ApJ 628, 604 (2005) 155 AMIGA: A New Model of Galaxy Formation and Evolution A. Manrique on behalf of the AMIGA collaboration Dpt. d’Astronomia i Meteorologia and Institut de Ciències del Cosmos, Univ. de Barcelona. C/ Martı́ i Franquès 1, E-08028, Barcelona, Spain, a.manrique@ub.edu Summary. Models of galaxy formation and evolution have become a fundamental tool to understand the observed properties of galaxies in a cosmological framework. They follow the evolution of the baryonic gas trapped in dark-matter halos as these grow hierarchically. Although they have improved during the last decade, current models still have limitations regarding the dynamical range and CPU time in numerical simulations or the resolution of merger trees in semi-analytical models. AMIGA is an analytical model based on an interpolation grid that includes all the relevant halo masses and reaches high redshifts. This feature allows one to study in a self-consistent way the coupled evolution of galaxies and inter-galactic medium (IGM) and the effects of a primordial stellar population (Pop III). The predictions of AMIGA are, in general, satisfactory and better than those of previous models. 1 Introduction Models of galaxy formation and evolution developed from the seminal paper of Frenk and White [9], and have become a fundamental tool to understand the observed properties of galaxies and how they change with time. The ultimate and fundamental goal of these models is to follow self-consistently through first principles the coupled evolution of dark matter and baryons from the primordial epoch, when both components grow linearly as density fluctuations, to the observation time. Unfortunately, there are relevant processes involved in galaxy formation and evolution that are poorly known, being star formation (and the related feedback effects) the most paradigmatic case. To deal with them, it is necessary to make assumptions and introduce free parameters. The essence of the fundamental goal can still be accomplished by modeling all the processes self-consistently, minimizing in this way the number of free parameters. Galaxy formation and evolution can be studied by means of two techniques: numerical simulations using N-body plus hydrodynamic codes, and (semi-)analytical models (SAMs). The first technique follow accurately the evolution of dark matter and baryons by solving the corresponding equations 158 Alberto Manrique (star formation, however, is treated approximately). Nonetheless, numerical simulations have a limited dynamical range and consume a lot of CPU time. They are appropriate to study the formation and evolution of individual galaxies or small groups. SAMs overcome the above limitations by modeling all the processes through physically motivated prescriptions. Although not as accurate as numerical simulations, they are more flexible and, therefore, more suitable to study the origin of the typical properties of galaxies. In this contribution I will focus on SAMs. 2 SAMs SAMs are made up of three basic components: the dark matter, which governs the dynamics, the baryons, which produces the observable radiation, and an interface to allow the comparison between the model output and observations. Dark matter fluctuations collapse and virialize to form halos. The evolution of dark matter halos is followed through a model of gravitational clustering, describing the mass growth history, and a model of the halo inner structure. Usually the latter model is not consistent at all with the gravitational clustering model. There are three key processes that drive the evolution of baryons: cooling, star formation and interactions. According to the general picture of galaxy formation, rotating hot gas, in hydrostatic equilibrium within the halo potential well, radiates and cools. Cooled gas falls towards the halo center conserving angular momentum and forms a disk where star formation takes place. As stars evolve they change their surroundings by ejecting material, metals and energetic radiation. These are the so-called feedback processes that influence the formation of the next generation of stars. Central disks become satellite galaxies when their host halos are engulfed by larger ones. Satellite orbits decay owing to dynamical friction, and they are captured by the central galaxy, causing its destruction and the formation of a bulge if the masses of the satellite and the central galaxy are comparable. SAMs use physically motivated prescriptions to model all these processes and also include a model of stellar population synthesis to account for the photometry and spectroscopy of galaxies. The latest versions comprise the effects of central super-massive black holes and use a radiative model for AGNs. The main groups that have contributed to the development of SAMs are those from Munich (Kauffmann and collaborators [12], [13]), Durham (Cole and collaborators, [5],[6] GALFORM), Santa Cruz (Somerville and Primack [15]) and Paris (Hatton and collaborators [10], GALICS). All these models predict galaxy properties (luminosity function, color-magnitude relation, TullyFisher relation, morphological fractions, disk sizes, etc.) that are in reasonable agreement with observations. This result implies that the modeling used by SAMs is also correct in general. However, a finer inspection reveals some problems. The first of them concerns the luminosity function (LF) at z = 0. SAMs AMIGA: A New Model of Galaxy Formation and Evolution 159 predict an excess of galaxies at “low” luminosities (i.e., L between the LF knee and the observational limit) and galaxies too luminous. The other two principal problems are related to the fact that SAMs use merger trees generated by Monte Carlo or N-body simulations to follow the growth history of halos. This technique causes limitations in the minimum halo mass and the initial epoch these models are able to deal with. For example, GALFORM, which employs Monte Carlo simulations, can only consider halo masses > 1010 M and redshifts z < 7. These limitations mean that galactic properties at (literally) low L or high z are unknown, as well as the effects of the adopted initial conditions. Figure 6 of [15] gives a nice summary of the results for the LF of the first batch of SAMs in a Standard CDM cosmology (Ωm = 1, h = 0.50) using different prescriptions for the star formation and stellar feedback. The deficiencies pointed above are evident in the B and K bands. Some mechanisms have been proposed to ameliorate the shape of the predicted LF. For instance, SN feedback helps correct the excess of galaxies at “low” luminosities, although it slightly extends the tail of bright galaxies. On the other hand, dust extinction notably cuts back the bright tail, but only in the B band. In the second batch of SAMs ([10], [2], [3]) alternative feedback processes have been considered in order to predict better LFs, this time for the concordance model (a flat, Ωm = 0.3, h = 0.70 universe). The existence of an ionizing background prevents that low-massive halos cool off. [2] obtain a “low” luminosity prediction close to observations by assuming that gas within halos with circular velocity < vcir ' 50 km/s cannot cool. A feedback process linked to AGNs has been suggested to avoid the formation of hyper-luminous galaxies. Although the physics involved is still quite speculative [4], it seems that a continuous injection of energy through the AGN jets may be capable of keeping the hot gas from cooling in massive halos. 3 AMIGA AMIGA (Analytical Model for IGM and GAlaxy evolution) is the first analytical model that follows the coupled evolution of galaxies and IGM in an almost fully self-consistent way. From the modeling viewpoint AMIGA differs from the previous SAMs in the following points: • Evolution of the inner structure of halos [14] (gas cooling, galaxy dynamics). Fully self-consistent and in good agreement with cosmological N-body simulations • Pop III (initial conditions for normal galaxy formation). Self-consistent except for three parameters: the metallicity separating Pop III from Pop I/II, the mass of Pop III stars, and the form of its IMF. • Galaxy-IGM interaction (ionizing background an associated effects). Selfconsistent except for the escaping fraction of ionizing photons from Pop I/II stars. 160 Alberto Manrique • Galaxy-galaxy and galaxy-halo interactions (star formation, galaxy dynamics and evolution). Self-consistent except for the halo truncation radius. • Detailed inner structure of galaxies (morphological properties). Fully selfconsistent for spirals, empirical for ellipticals. • Coupled evolution of super-massive black holes and galaxies (quasars, dynamical and feedback effects on galaxies). Fully self-consistent. AMIGA does not use merger trees generated by Monte Carlo or N-body simulations to follow the evolution of halos. Instead it applies an interpolation grid in halo masses and redshifts where it stores the halo properties (including those of the galaxies they host). The strategy followed to build this grid is sketched in Fig. 1. Its implementation allows AMIGA to reach high redshifts (zini ' 60 in our calculations in the concordance model) and improve the computation of the galaxy/IGM interaction (since the properties of all halos are known at each z). However, the construction of the grid requires large amounts of memory and CPU time (the latter limitation can be overcome by parallelizing the code). For example, if the observed properties of galaxies are to be predicted with a resolution of 1 magnitude, the code needs more than 8 GB of memory and more than 20 days of CPU. progenitors z z z max obs M halo at formation new point accretion track trivial properties low M M up Fig. 1. Interpolation grid. The halo properties at a given z are computed from the progenitor properties at the formation time and their evolution tracked to the current epoch AMIGA: A New Model of Galaxy Formation and Evolution 161 4 Results Since AMIGA deals with intrinsic luminosities (the current version does not include any model of dust extinction) the results shown in this section refers to the K band, which makes easier the comparison with empirical data. I will focus on the LF because, in addition to being a fundamental property of galaxy population, it is well determined over a restricted luminosity range. 4.1 The Classical Solution Using the standard galactic parameters, in particular a star formation efficiency 0 = 0.03 and a sharp cutoff for the minimum circular velocity vmin = 47 km/s, AMIGA predicts a LF similar to that of precedent SAMs. As shown in Figure 2 the agreement with empirical LFs is fair at “low” L, but still there are some galaxies too luminous. This result, which I will refer to as the classical solution, indicates that initial conditions (PopIII, cosmic reionization) plays no role on the present LF, at list over the observed luminosity range. Consequently, it validates the predictions made by previous SAMs, despite of not accounting for initial conditions. The classical solution rises the following three questions: • Can the prediction of the LF bright tail be improved? • Can a self-consistent treatment of the photo-ionization feedback keep the good agreement obtained for the “low” luminosity range? This is a relevant issue, since the the shape of the LF at “low” L is sensitive to the value of the parameter vmin . • How do galaxies look like at low luminosities or high redshifts? The features of AMIGA make this model suitable to deal with the above questions, and let one go one step beyond by considering its predictions on dwarf galaxies and galaxies at high redshift. Both predictions can be checked out simultaneously through one empirical quantity: the galaxy number counts. 4.2 New Results Instead of resorting to the AGN feedback, a variable star formation efficiency depending on the galaxy mass, Mgal is introduced to ameliorate the shape of the LF at high L: !−α Mgal (Mgal ) = 0 (1) 0 Mgal 0 = 8 × 109 M , and α = 0.8. The left panel of Figure 3 show the where Mgal predicted LF in this case. According to equation (1), massive galaxies form stars less efficiently than low-mass galaxies. This fact makes the LF bright 162 Alberto Manrique Fig. 2. Comparison of the predicted LF in the K band for the classical solution with three empirical LFs [11], [1] . Notice the difficulty of determining observationally the LF over the “low” L range. The arrow marks the maximum observed magnitude Fig. 3. Left panel: Predicted LF in the K band using a variable star formation efficiency (eq. [1]) and a self-consistent method to model the photo-ionization feedback. Right panel: Extension of the predicted LF to low L using three prescriptions for the photo-ionization feedback AMIGA: A New Model of Galaxy Formation and Evolution 163 tail fall faster and be closer to observations when comparing with the classical solution, but it also produces an excess of cold gas at z = 0. The left panel of figure 3 also shows the predicted LF when dealing with the photo-ionization feedback in a self-consistent way (thanks to the galaxyIGM interaction implemented in AMIGA). The plot confirms that the ionizing background causes the same effect on the “low” luminosity range than the ad hoc parameter vmin . AMIGA gives a robust result for this range that makes one wonder about the predicted LF at very low luminosities (within the realm of dwarf galaxies). This prediction is displayed in the right panel of figure 3, and compared with those using a sharp vmin , as in the classical solution, and a variable vmin , as suggested by the theoretical work [8], which considers the evolution of the ionizing flux with time. Up to magnitudes brighter than −10, the self-consistent result is compatible with the (no self-consistent) model [8], while the result arising from the sharp vmin decreases faster for increasing magnitude. Fig. 4. Left panel: Cosmic histories for the star formation rate, gas and star masses, ionizing flux and ionized fraction. Right panel: Comparison of the predicted galaxy number counts in the K band with the empirical determination of [7] AMIGA predicts values of global properties of the universe (stellar formation rate, cold gas density, epoch of reionization) at z = 3 in fair agreement with observations. This result grants confidence to the model predictions for galaxies at high redshift. However, not only does AMIGA give values of cosmic properties at z = 3, but also their whole history. The left panel of figure 4 shows some examples of cosmic histories. Galaxy number counts in magnitude involve galaxies of different luminosities at different redshifts. The zero-point of galaxy counts depends basically 164 Alberto Manrique on the bright tail of the LF, while the slope depends on the evolution of the LF (and also on the expansion of the universe). If the shape of LF does not change a constant slope equal to 0.6 develops. The right panel of figure 4 show the predictions of AMIGA for the self-consistent case and the classical solution. The self-consistent case gives a good prediction up to mk ' 19 but it deviates at larger magnitudes, which corresponds to dim galaxies formed at z ' 1.5 5 Conclusions AMIGA is the first model that accurately follows the coupled evolution of galaxies and IGM from the dark ages in an almost fully self-consistent way. The results regarding the history of the universe at high redshift are in reasonable agreement with the few empirical data available. The present LF within the observed range does not depend on the initial conditions at high redshift. AMIGA predicts a correct LF in the K band at high L, but using an ad hoc star formation efficiency that causes an excess of cold gas at z = 0. It also predicts a correct LF at “low” L, without resorting to ad hoc approaches for the photo-ionization feedback, and, for the first time, predicts the LF for dwarf galaxies. References 1. E. F. Bell, D. H. McIntosh, N. Katz & M. D. Weinberg: ApJSS 149, 289 (2003) 2. A. J. Benson, R. G. Bower, C. S. Frenk, C. G. Lacey C. M. Baugh, & S. Cole: ApJ 599, 38 (2003) 3. R. G. Bower, A. J. Benson, R. Malbon, J. C. Helly, C. S. Frenk,C. M. Baugh, S. Cole, & C. G. Lacey: MNRAS 370, 645 (2006) 4. A. Cattaneo, A. Dekel, J. Devriendt, B. Guiderdoni, & J. Blaizot: MNRAS 370, 1651 (2006) 5. S. Cole, A. Aragon-Salamanca, C. S. Frenk, J. F. Navarro, & S. E. Zepf: MNRAS 271, 781 (1994) 6. S. Cole, C. G. Lacey, C. M. Baugh, & C. S. Frenk: MNRAS 319, 168 (2000) 7. D. Cristóbal-Hornillos, et al.: ApJ 595, 71 (2003) 8. M. Dijkstra, Z. Haiman, M. J. Rees, & D. H. Weinberg: ApJ 601, 666 (2004) 9. C. S. Frenk, S. D. M. White: ApJ 61, 33 (1991) 10. S. Hatton, J. E. G. Devriendt, S. Ninin, F. R. Bouchet, B. Guiderdoni, & D. Vibert: MNRAS 343, 75 (2003) 11. J.-S. Huang, K. Glazebrook, L. L. Cowwie, & C. Tinney: ApJ 584, 203 (2003) 12. G. Kauffmann, S. D. M. White, & B. Guiderdoni: MNRAS 264, 201 (1993) 13. G. Kauffmann, J. M. Colberg, A. Diaferio & S. D. M. White: MNRAS 303, 188 (1999) 14. E. Salvador-Solé, A. Manrique, & J M. Solanes: MNRAS 358, 901 (2005) 15. R. S. Somerville ,& J. R. Primack: MNRAS 310, 1087 (1999) The innermost regions of Active Galactic Nuclei – from radio to X-rays E. Ros1 , M. Kadler2,? , S. Kaufmann3 , Y.Y. Kovalev1,4, J. Tueller2 , and K.A. Weaver2 1 2 3 4 Max-Planck-Institut für Radioastronomie, Auf dem Hügel 69, D-53121 Bonn, Germany, ros@mpifr-bonn.mpg.de Astrophysics Science Division, Code 662, NASA’s Goddard Space Flight Center, Greenbelt Road, Greenbelt, MD 20771, USA mkadler, tueller, kweaver@milkyway.gsfc.nasa.gov Argelander-Institut für Astronomie, University of Bonn, Auf dem Hügel 71, D-53121 Bonn, Germany kaufmann@astro.uni-bonn.de Astro Space Center, P. N. Lebedev Physical Institute, ulitsa Profsoyuznaya 84/32, 117997 Moscow, Russia ykovalev@mpifr-bonn.mpg.de Summary. Active Galactic Nuclei can be probed by at different regions of the electromagnetic spectrum: e.g., radio observations reveal the nature of their relativistic jets and their magnetic fields, and complementarily, X-ray observations give insight into the changes in the accretion disk flows. Here we present an overview over the AGN research and results from an ongoing multi-band campaign on the active galaxy NGC 1052. Beyond these studies, we address the latest technical developments and its impact in the AGN field: the Square Kilometre Array, a new radio interferometer planned for the next decade, and the oncoming X-ray and gamma-ray missions. 1 Background The standard model for Active Galactic Nuclei (AGN) proposes that the energy release is produced by the accretion of matter onto super-massive black holes (BH) [11, 44, 45]. The AGN is powered by the conversion of gravitational potential energy into radiation, although the rotational kinetic energy of the BH may also serve as an important source of energy [36, 64, 37]. A fraction of the matter is ejected via a poloidal magnetic field in a jet perpendicular to the accretion disk surrounding the black hole. A region of gas with broad emission lines is located close to the accretion disk. Narrow-line emitting clouds are present outside the disk and torus region. AGN unification models [1] presume that depending on the viewing angle of the torus-disk-jet complex the observed galaxy appears as a blazar when the ? NASA Postdoctoral Research Associate 166 Ros, Kadler, Kaufmann et al. jet points towards the observer; as an object like Seyfert 1, Broad-Line-Radio Galaxy or Quasi-Stellar Object for intermediate angles; or as a Seyfert 2 or a Narrow-Line-Radio Galaxy when the jet lies in the plane of the sky. For a phenomenological taxonomy of the AGN zoo, see e.g. Table 1.2 in [40]. The powerful jets observed commonly in radio-loud AGN [74] consist of relativistic (shocked) plasma which may extend up to kiloparsec scales (showing typically extended radio lobes [56, 6]), much larger in size than their host galaxies. Jets oriented close to the line of sight have favourable observing conditions due to relativistic boosting. Emission from AGN can be observed throughout the electromagnetic spectrum: The jet synchrotron emission can be probed by radio and millimetre observations. The brightest jets emit also in the optical (e.g., [3]) and in X-rays [17]. The thermal emission of the accretion disk and the surrounding torus are probed both in temperature distribution and in morphology by infrared interferometry (e.g., [23]). The broad and narrow emission line regions are studied by optical spectroscopy. X-ray imaging and spectroscopy probe the corona, the accretion disk, and the jet radio lobe region at kiloparsec-scales as well as compact jets in blazars.. Shocked regions in the jet, where the plasma is hotter, can emit at energies of up to γ-rays [55]. Presently, different tools are available for the astronomers to probe the nature of AGN. The spectral energy distribution can be split into several components produced by distinct emission mechanisms (synchrotron, inverse Compton, thermal, etc.) and affected by absorption. Spectroscopy of the Fe Kα X-ray emission (the strongest fluorescent line due to the highest cross section for the absorption of all iron atoms less ionised than Fe+16 ) probes the relativistic accretion disk. This line was first detected in radio-quiet (Seyfert 1 type) galaxies (e.g., MCG–6-30-15 [68]). The “louder” the galaxy is at radio wavelengths, the weaker the iron line tends to be. A thermal “bump” is usually present in the optical and ultraviolet continuum spectrum. Spectroscopy of the broad and the narrow line regions provides information about the pressure of the medium around the jet. Measurements of the variability of radio sources yield limits on the size of the emitting regions (from the smallest timescales of variations). Radio- and millimetre-wave imaging at the highest resolutions (very-long-baseline interferometry: VLBI) provide resolutions down to 0.1 milli-arcsec, reaching typically sub-parsec scales. This shows the jet structure at the innermost region of the AGN. At the highest frequencies, the emission from the jet base (core) is unveiled [41, 43]. Measurements of the polarisation reveal changes in the magnetic fields present at the jet. Single-dish flux-density and spectral monitoring programs probe absorption effects and the presence of different synchrotron-emitting features in the jets. These are complemented by X-ray monitoring to probe the accretion disk via spectroscopy and imaging. These aspects will be expanded in the next sections. First we give an overview on the extensive work being performed on AGN at different wavelengths, then we will describe an ongoing campaign on the active galaxy NGC 1052, and finally we will provide some prospective view AGN from radio to X-rays 167 of future observations with the Square Kilometre Array and new X- and γ-ray missions. 2 Observing the multi-waveband sky Pioneering work combining VLBI and X-ray observations was performed already in the 1970s [47]. Important landmarks in this research are, for instance, the combined radio and X-ray observations on the quasars 3C 120 [48] —where the X-ray flux drops for days to weeks just prior to the ejection of bright features in the jet, and PKS 1510–089 —where a superluminal ejection in the jet occurred immediately after the start of a major X-ray and optical outburst in late 2000 [49]. AGN surveys are an essential tool for finding and identifying appropriate candidates for successful combined X-ray and radio studies. Ref. [44] summarises most of the ongoing surveys in VLBI, radio monitoring, infra-red, optical, X-ray and γ-ray wavelengths. VLBI is a well-established technique. Four decades have elapsed since the first experiments (e.g., [32]), and the discovery of superluminal motions [73, 8] took place thirty years ago. The exploration of the radio sky on parsecscales has been facilitated dramatically since the construction of the Very Long Baseline Array (VLBA) in the mid 1990s. Starting with the regular observations of the VLBA a program to monitor the jet kinematics of the most prominent radio-loud AGN (over hundred) in the northern sky was defined and initiated in 1994. This was the 2 cm Survey [33, 75, 35, 39], continued since 2002 as the MOJAVE program [42, 20] —the latter including also linear and circular polarisation monitoring. X-ray astronomy is also a relatively young science, experiencing a new revolution with every new generation of X-ray missions. After ASCA (1993– 2001, [67]) and BeppoSAX (1996-2003, [5]), the missions Chandra (since 1999 [72]), XMM-Newton (since 2001 [25]), and Suzaku (since 2005 [54]) constitute the state of the art for X-ray imaging and spectroscopy at present. X-ray emission from the sources of the 2 cm Survey and the MOJAVE samples has been studied in detail from the available archival data [28]: 2cmX-sample, was established by making use of all publicly available archival data from the first four missions mentioned above. Originally with 50 sources, the sample is being completed by a Swift program to observe the remaining 83 objects from the MOJAVE sample [29]. In the following section we report on the multi-wavelength monitoring campaign on one particular source from the 2cm-X-Sample, the active galaxy NGC 1052. 2.1 NGC 1052: the key to jet-disk coupling The nearby elliptical galaxy NGC 1052 can be classified as a radio-loud object [27]. It hosts a twin-jet system oriented close to the plane of the sky (e.g., 168 Ros, Kadler, Kaufmann et al. [70]). NGC 1052 has been classified as the prototypical low-ionisation nuclear emission region (LINER). This source is particularly suited for connecting radio and X-ray observations, since it shows an edge-on accretion disk, water maser emission, an obscuring torus (see below), and it hosts mildly relativistic jets that can be probed by VLBI. Detailed multi-wavelength observations at the centimetre range provide evidence for an obscuring torus covering partially the western, receding jet [34, 31, 27]. The column densities measured in the radio and X-rays have comparable values [26]. In X-rays the source shows a flat spectrum and a soft excess [71, 15, 16, 26]. Sub-parsec imaging of both jets by VLBI from the 2 cm Survey/MOJAVE programme [70] revealed motions of 0.26 c. A new jet feature is ejected every 3–6 months, correlated with flux density outbursts. There are indications [28] that the ejection of a new feature in the jet, estimated to occur in Epoch 2001.0, is associated with the change of the relativistic line profile from data taken by BeppoSAX ([16], reanalysed in [28]) at epoch 2000.03 and by XMM-Newton at epoch 2001.62, where the line is broadened. This is the first detection of a highly relativistic iron line in a radio-loud AGN with a compact radio jet. The variability of the iron line and of the fraction of accreted energy that is channeled into the jet could be related to changes in the structure of the magnetic field in and above the accretion disk. Given this scenario, we initiated in mid 2005 a multi mission campaign to track the birth of new VLBI components at the base of the jet and counter-jet, to compare those with the flux density monitoring and spectroscopy in radio and X-rays and to establish cause-effect relationships in a much more confident way than the accretion-ejection event reported previously [28]. This campaign includes, in X-rays: a) Rossi X-Ray Timing Explorer (RXTE) flux density monitoring at 2-10 keV: 30 epochs of 2 ks each, scheduled every three weeks; b) Chandra imaging and spectroscopy: one deep observation in Sep 2005; c) XMM-Newton imaging and spectroscopy: one triggered observation in Feb 2006 so far. The source is also being monitored by the Burst Alert Telescope (BAT; see [4]) on-board Swift since the beginning of 2005 (see below). Radio observations include: a) λλ 13/6/3.6/2.8/2/1.3/0.9cm dedicated light curves taken by the 100-m radio telescope in Effelsberg, with ca 70 h observations scheduled every three weeks; b) λλ 31/13/7.7/6/3.9/3.6/2.7/2/1.4cm light curves taken by the RATAN-600 and the Univ. of Michigan Radio Astron. Obs. (UMRAO, [2]) in the framework of long-term monitoring programs; and c) λλ 13/7mm Very Long Baseline Array imaging, with 18 observing runs of 6 h each scheduled every six weeks (images from the first epochs are presented in [62]). RXTE Monitoring The Rossi X-ray Timing Explorer (RXTE) is monitoring NGC 1052 since mid 2005 with pointings of 2 ksec each every three weeks. We concentrate on the analysis of the data from the PCA detector [24]. The AGN from radio to X-rays 169 data were reduced with the rex script1 that simplifies the reduction of large amount of data, with standard criteria for faint sources. We used the data of PCU 2 and layer 1 which provide the best signal-to-noise ratio. For the spectral analysis we used xspec version 11.3. We restricted the analysis to the energy range 2–10 keV and fitted an absorbed power law with a fixed value for the Galactic absorption of NH = 2.95 · 10−20 cm−2 [30]. Fig. 1. Results from the multi-mission campaign on NGC 1052. a) RXTE monitoring results, showing the 2–10 keV flux (top) and photon index (bottom) as a function of time. b) Hard X-ray image in the 15–150 keV band. In this ∼ 5◦ × 5◦ image, only three sources at > 5σ are present; c) Hard X-ray light curve from the first 16 months of BAT observations. d) & e) Selected RATAN-600 instantaneous continuum spectra over the last 10 years of observations (d) and during the time of our multi-frequency campaign (e). The evolution of new radio flares resulting from parsec-scale jet components is clearly visible in the overall spectrum changes. The RXTE light curve (2–10 keV) and X-ray spectral evolution of NGC 1052 is shown in Fig. 1 a). Bearing in mind the complex nature of the X-ray spectrum of NGC 1052 [28], the characterisation of the RXTE data from the individual scans with a simple power-law model requires some extra care in the interpretation. While the flux in the 2–10 keV band is relatively insensitive to imperfect spectral modeling, changes of the formal spectral index can be due to changes of the ratio between various components contributing to the 1 See http://heasarc.gsfc.nasa.gov/docs/xte/recipes/rex.html 170 Ros, Kadler, Kaufmann et al. spectrum (e.g., the soft excess, and the reflection component), as well as due to changes of the primary power-law continuum component or the absorption. A more careful analysis of the RXTE data with the detailed spectral composition coming under scrutiny from the deep pointings of XMM-Newton and Chandra is underway. The RXTE monitoring data show a systematic increase of flux mid to end 2005 by more than a factor of 2 followed by a dramatic drop around epoch 2006.1. After that, the source flux rises again over several months. With the present data, it is difficult to judge whether this sampled portion of the light curve corresponds to two outbursts. Alternatively, the data could be interpreted as showing two long dips similar to the ones in 3C 120 [49] but on longer time scales. The photon index Γ is found to vary most of the time between 1.4 and 1.8, values that are typically seen in Seyfert galaxies. It is interesting to note that the historically well-known “unusually flat X-ray spectrum” of NGC 1052 [71, 16, 26, 27] is found only during two relatively short time periods in Sep/Oct 2005 and Feb/Mar 2006. These both epochs coincide with the beginning of a rise in the X-ray flux, indicating that either flares occur first at high energies, or vice-versa that the dips last longest in the soft band. Swift BAT Monitoring Figure 1 b) shows the hard X-ray image (15–150 keV) of NGC 1052 from the first 16 months of Swift BAT monitoring observations [46]. The source is clearly detected with a formal significance of >5σ. Unlike the lower-energy bands (radio, soft–medium X-rays), the BAT light curve (Fig. 1 c) ) shows only marginal variability on time scales of months at hard X-rays. Note that negative count rates can statistically result from the subtraction of two almost equal numbers in the background-dominated limit and that large error bars indicate periods of low exposure in the region of NGC 1052 on the sky. For these two reasons, the negative value at ∼ 2005.6 should not be over-interpreted. The perhaps more significant feature is the decrease and subsequent increase in the first three months of 2005. From mid through end 2005, the hard X-ray flux of NGC 1052 was quasi-constant within the sensitivity limit of BAT. The monitoring of NGC 1052 at hard X-rays will continue throughout the regular all-sky observations of BAT. This represents a further valuable component in our monitoring campaign of this source. Both long-term trends and putative higher-amplitude variability on shorter time scales (e.g., due to SSC flares) will be detectable and can be analysed in view of the variability patterns at lower energies. In particular, it shall be noted that BAT is sensitive enough and that NGC 1052 is bright enough in the 15–150 keV band to detect variability if it occurs with the same amplitude as in the RXTE band. Thus, even a lack of variability through 2006 in the BAT light curve would put an important constraint on the nature of the nuclear activity in NGC 1052 by attributing most of the variability to a spectral component at soft X-ray energies. AGN from radio to X-rays 171 RATAN-600 Radio Spectra We started the long-term monitoring of 1–22 GHz radio spectra of NGC 1052 at the 600 meter ring radio telescope RATAN600 of the Russian Academy of Sciences in 1996. The observations of continuum spectra are done almost instantaneously at 1, 2.3, 3.9/4.8, 7.7, 11.1, and 22 GHz in a transit mode, 2–4 epochs per year. Details of the observations and data processing can be found in Ref. [38]. Selected RATAN-600 instantaneous continuum spectra over the last 10 years of observations as well as the data accumulated during the time of our multi-frequency campaign until mid 2006 are presented in Figure 1 c) & d). The flux density at frequencies over 10 GHz has dropped by a factor of 2 since the start of the campaign. The observed flaring variability is most probably due to production and evolution of new parsec-scale features dominating in the synchrotron spectrum of the compact jet. The flare spectrum peaks around 10 GHz in 2005, with SSA radiation below the peak. The characteristics timescale of the observed long-term radio variability of NGC 1052 is several months. Summary of Some Early Results from the NGC 1052 Monitoring Campaign The monitoring campaign of NGC 1052 begun in mid 2005 and is en route to a scrutiny of the jet-disk coupling in this active galaxy. For one-and-a-half years now, the jet-production activity has been monitored at sub-milli-arcsecond angular resolution with the VLBA at 22 and 43 GHz [62]. Over this period, the source was constantly active in the radio. Its radio spectrum has been monitored every three weeks in the cm regime with the Effelsberg 100-m telescope, and within the long-term monitoring programs at the University of Michigan and with RATAN-600. We are in the process of making a detailed analysis of these data. This will enable us to compare the jet-production activity to accretion-disk probing observations at high energies. Our preliminary analysis of the first 1.5 years of RXTE monitoring data reveals for the first time the previously missing piece of evidence that the 2–10 keV X-ray spectrum of NGC 1052 is variable on essentially the same time scales as the radio emission. While a continuation of the monitoring over several variability cycles is important for quantifying this finding. The present data support the idea that the X-ray and radio components may be directly (or indirectly) coupled. In this context, it is important to consider our earlier finding that the iron-line in NGC 1052 has varied along with a major jet-ejection event [28] on exactly these time scales in 2000/2001. Additional deep X-ray spectroscopic observations with XMM-Newton and/or Suzaku will likely be able to find very different X-ray spectral states in terms of both the continuum and the iron-line and can be interpreted in view of the continuously monitored jet-production activity. In particular Suzaku, with its high effective area at ∼ 6 keV and its broad band pass that covers also the hard X-ray regime, will yield important constraints on the iron line and will at the same time be able to reconcile the RXTE and XMM-Newton results with the Swift BAT results. 172 Ros, Kadler, Kaufmann et al. 3 Instruments for the future In this section we discuss the prospectives for AGN research in the radio and X-rays under the light of the astronomical facilities planed for this and the next decade. Radio: the Square Kilometre Array Two new radio facilities will be available in the near future: operating at very long wavelengths, the Low Frequency Array in The Netherlands (LOFAR, [61]), and the Atacama Large Millimetre Array in Chile (ALMA, [7]) at the sub-millimetre range. In the cm-band, the Square Kilometre Array (SKA, [63, 69, 9]), a new facility reaching 100 times better sensitivity at milli-arcsec-resolution is planned for the next decade. The scientific case of the SKA requires a radio telescope with sensitivity to detect and image atomic hydrogen at the edge of the universe, which requires a very large collecting area. The new instrument should have a fast surveying capability over the whole sky, which makes a very large angular field of view mandatory. It should have capability for detailed imaging of the structures at the sub-arcsecond level, for which a large physical extent is needed. Finally, a wide frequency range is needed to handle the different scientific goals. The concept which fills these requirements implies a square kilometre collecting area in an interferometer array, with a sensitivity two orders of magnitude larger and a survey speed four orders of magnitude larger than the Expanded Very Large Array. The proposed frequency range should cover from 0.1 to 25 GHz, with baselines up to 3000 km, and a field-of-view of 50 square degrees at frequencies lower than 1 GHz. This project would cost over one billion euros with a running cost of around seventy million euros per year. The socalled Phase 1 of the project should be ready in 2012, and the complete array at the end of the 2010s. The short list of site candidates favours a location in Southern Africa or Western Australia. A reference design [22] has been developed recently, including three system types to be combined in hybrid elements. The reference design includes a sparse aperture array (0.1–0.3 GHz, a.k.a. as “Era of Recombination” array, similar to LOFAR, providing wide and multiple independent field-of-views), a dense aperture array (0.3–1.GHz, a.k.a. radio “fish eye” lense with all-sky monitoring capability), and a smalldish and “smart-feed” array (0.3–25 GHz, a.k.a. “radio camera”, with ∼10-m dishes and wide response feeds). More information on the project can be found under http://www.skatelescope.org. The SKA will image all radio galaxies in the sky to the micro-jansky level, especially probing active galaxies in the radio-quiet regime (e.g., the Seyfert galaxies exhibiting broadened iron lines in X-Rays). The project will signify a revolution in observational cosmology, in the studies of the magnetic universe, etc., but will also enable a completely new view of the traditional targets of VLBI research: active galactic nuclei. High-energy missions Research at high energies will reach new frontiers in coming years with the missions currently being put into operation or planned (see [57] for a recent review on X-ray missions): AGN from radio to X-rays 173 Suzaku (since 2005) is specifically designed to study the Fe Kα line. Its broad band pass from 0.3 keV to > 100 keV combined with its high effective area at ∼ 6 keV and the good spectral resolution makes it possible to determine the continuum model and at the same time measure the iron K line. A recent review of Suzaku observations of iron lines in AGN can be found in [60]. Within its blazar key project, the X-Ray Telescope (XRT) of Swift is currently obtaining a large number of X-ray spectra of radio-loud, core-dominated AGN, many of which have never been observed in the soft X-ray regime since ROSAT in the 90s and many never above 2 keV. In particular, Swift is going to complete the 2 cm-X-Sample of X-ray observed MOJAVE sources [28], the 133 radio-brightest compact AGN in the northern sky. Swift is particularly well suited for such a large pointed-survey program because of its flexibility. It will greatly enhance our knowledge of radio-loud AGN X-ray spectra by providing a VLBI-defined statistically complete set of X-ray (and UV) spectra. This will be particularly valuable to identify more sources similar to NGC 1052 for combined VLBI and high-energy studies. The next planned major step for X-ray observations of AGN will be Constellation-X [21]. The current mission design of Con-X foresees a sensitivity of > 50 times better than XMM-Newton and Suzaku. The mission will provide the highest spectral resolution to date by making use of calorimeter detectors. These will be particularly important for fine-scale studies of relativistic broad iron lines in AGN. With Con-X, it will be possible to obtain high-quality X-ray spectra of accreting super-massive black-hole systems in short snapshot observations. For sources like NGC 1052, this means that it will be possible to perform high-sensitivity accretion-disk monitoring with a minimum of required telescope time. At the other end of the electromagnetic spectrum, the SKA will for the first time provide the possibility to study in detail the radio cores of Seyfert galaxies and other radio-quiet AGN. Together, Constellation-X and the SKA will allow us to perform high-sensitivity combined radio and X-ray studies of AGN, both radio-loud and radio-quiet. In the hard X-ray regime, the all-sky monitoring Energetic X-ray Imaging Survey Telescope (EXIST, [14]) would yield a sensitivity a factor 50–100 higher than Swift/BAT at the 5–600 keV. The science goal of EXIST is the discovery and study of black holes on all scales from stellar-mass to super-massive black holes. In the context of hard X-ray blazar studies, EXIST is expected to boost the number of observationally accessible sources (e.g., only about 10% of the MOJAVE blazars are bright enough to be detected by BAT after 16 months of observations, while we expect most if not all MOJAVE sources to be easily detectable by EXIST). For the study of sources like NGC 1052 that are bright enough to be studied already today by the BAT (see above), EXIST will dramatically increase the time resolution and decrease the minimal detectable variability amplitudes. At even higher photon energies, the new γ-ray facility GLAST (Gamma-ray Large Area Space Telescope; [13]) will be launched in late 2007. It follows in 174 Ros, Kadler, Kaufmann et al. the footsteps of the Compton Gamma Ray Observatory (CGRO; 1991-1999) whose main instrument EGRET has discovered that blazars and flat spectrum radio quasars are strong γ-ray emitters [12]. The third EGRET catalog of highenergy γ-ray sources [18] originally contained 66 high-confidence identifications with blazars. Among the statistically complete MOJAVE sample of the 133 radio-brightest compact AGN in the northern sky, 44 have a high-probability EGRET identification according to revisions of the EGRET-blazar sample [51, 52, 65, 66]). While the detailed mechanism for the production of gamma-ray emission has not yet been agreed upon, it is widely accepted that the γ-ray emission from blazars is highly beamed and anisotropic [10]. It probably takes place close to the central engine or at the base of the relativistic jet. VLBI observations provide the best imaging probes close to the central engine, so that one expects to find differences in the milliarcsecond-scale radio properties and kinematics between EGRET-detected and not-detected sources. Indeed, [35] find that radio jets that are also strong γ-ray sources have faster jets than EGRET undetected sources. Furthermore, recent work [39, 42] indicates that EGRET sources have more compact radio cores and more highly polarised and more luminous jet features than non-EGRET sources. In the GLAST era, it will be possible to investigate these findings in unprecedented detail. The EGRET counterpart on-board GLAST will be the LAT (Large Area Telescope) whose capabilities are specifically well-suited for blazar-variability studies (e.g., [53]). The highly superior sensitivity and angular resolution of LAT is expected to result in the detection of thousands of new γ-ray sources, in particular it is expected that the LAT will be able to monitor the γ-ray light curves of hundreds of bright blazars with high temporal resolution. Various efforts in the radio regime are being made to exploit the opportunities offered by this mission: in particular, within the MOJAVE project, it is planned to trigger VLBI monitoring observations by GLAST/LAT detected γ-ray flares, to relate the flaring activity with ejections in the radio jet. It will also be possible to use GLAST/LAT light curves as jet-activity monitors. For sources like 3C 120 or NGC 1052, whose mass-accretion can be monitored via X-ray observations, this will open a new avenue to jet-formation studies. Acknowledgements: MK was supported by a NASA Postdoctoral Program Fellowship appointment conducted at the Goddard Space Flight Center. YYK is a Research Fellow of the Alexander von Humboldt Foundation. RATAN–600 observations were supported partly by the NASA JURRISS program (W–19611) and the Russian Foundation for Basic Research (01–02–16812, 02–02–16305, 05–02–17377). The campaign of observations of NGC 1052 is being performed in the framework of a wide collaboration including the 2cm Survey and the MOJAVE teams, and also particularly E. Ros, E. Angelakis, A. Kraus, Y.Y. Kovalev, A.P. Lobanov and J.A. Zensus at the MPIfR; J. Kerp and S. Kaufmann at the AIfA of the University of Bonn; M. Kadler, J. Tueller and K. Weaver at NASA/GSFC; A.P. Marscher at AGN from radio to X-rays 175 Boston University; and H.D. Aller, M.F. Aller and J. Irwin at the Univ. of Michigan. We thank M. Perucho and A.P. Lobanov for useful comments to the manuscript. References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. R. Antonucci: ARA&A 31, 473 (1993) M.F. Aller, H.D. Aller, P.A. Hughes: ASP Conf. Ser. 300, 159 (2003) J.N. Bahcall, S. Kirhakos, D.P. Schneider et al: ApJ 452, L91 (1995) S.D. Barthelmy, L.M. Barbier, J.R. Cummings et al: Space Science Rev. 120, 143 (2005) G. Boella, R.C. Butler, G.C. Perola et al: A&AS 122, 299 (1997) A.H. Bridle, R.A. Perley: ARA&A 22, 319 (1984) R.L. Brown, W. Wild, C. Cunningham: Adv. Space Research 34, 555 (2004) M.H. Cohen, W. Cannon, G.H. Purcell et al: ApJ 170, 207 (1971) J.M. Cordes: BAAS 208, 73.02 (2006) C.D. Dermer, R. Schlickeiser: ApJS 90, 945 (1994) A.C. Fabian: PNAS 96, 4794 (1999) C.E. Fichtel: ApJS 90, 917 (1994) N. Gehrels, P. Michelson: APh 11, 277 (1999) J.E. Grindlay, W.W. Craig, N.A. Gehrels et al: Proc. SPIE 4851, 331 (2003) M. Guainazzi, L.A. Antonelli: MNRAS 304, L15 (1999) M. Guainazzi, T. Oosterbroek, L.A. Antonelli, G. Matt: A&A 364, L80 (2000) D.E. Harris, H. Krawczynski: ARA&A 44, 463 (2006) R.C. Hartman, D.L. Bertsch, S.D. Bloom et al: ApJS 123, 79 (1999) J.F. Helmboldt, G.B. Taylor, S. Tremblay et al: AJ, submitted (2006) D.C. Homan, M.L. Lister: AJ 131, 1262 (2006) A.E. Hornschemeier, N.E. White, H. Tananbaum: AIP Conf. Proc. 774, 383 (2005) International SKA Project Office, Memo 69 http://www.skatelescope.org/PDF/memos/69_ISPO.pdf (2006) W. Jaffe, K. Meisenheimer, H.J.A. Röttgering et al: Nature 429, 47 (2004) K. Jahoda, J.H. Swank, A.B. Giles et al: Proc. SPIE 2808, 59 (1996) F. Jansen, D. Lumb, B. Altieri et al: A&A 365, L1 (2001) M. Kadler, J. Kerp, E. Ros et al: A&A 420, 467 (2004) M. Kadler, E. Ros, A.P. Lobanov et al: A&A 426, 481 (2004) M. Kadler: Compact Radio Cores in AGN: The X-Ray Connection. Dissertation, Rheinische Friedrich-Wilhelms-Universität zu Bonn (2005) M. Kadler et al: in preparation P.M.V. Kalberla, W.B. Burton, D. Hartmann et al: A&A 440, 775 (2005) S. Kameno, S. Sawada-Satoh, M. Inoue et al: PASJ 53, 169 (2001) K.I. Kellermann, B.G. Clark, C.C. Bare et al: ApJ 153, L209 (1968) K.I. Kellermann, R.C. Vermeulen, J.A. Zensus, M.H. Cohen: AJ 115, 1295 (1998) K.I. Kellermann, R.C. Vermeulen, M.H. Cohen, J.A. Zensus: BAAS 31, 856 (1999) K.I. Kellermann, M.L. Lister, D.C. Homan et al: ApJ 609, 539 (2004) S. Koide, K. Shibata, T. Kudohy et al: Science 295, 1688 (2002) S.S. Komissarov: MNRAS 359, 801 (2005) 176 Ros, Kadler, Kaufmann et al. 38. Y.Y. Kovalev, N.A. Nizhelsky, Yu.A. Kovalev t al: A&AS 139, 545 (1999) 39. Y.Y. Kovalev, K.I. Kellermann, M.L. Lister et al: AJ 130, 2473 (2005) 40. J.H. Krolik: Active Galactic Nuclei (Princeton Univ. Press, Princeton, NJ 1999), p 20 41. S.S. Lee, A.P. Lobanov, T.P. Krichbaum et al. In: 8th EVN Symp., Proc. of Sci., in press (2006) [arXiv:astro-ph/0611308] 42. M.L. Lister, D.C. Homan: AJ 130, 1389 (2005) 43. A.P. Lobanov, T.P. Krichbaum, D.A. Graham et al: A&A 364, 391 (2000) 44. A.P. Lobanov, J.A. Zensus. In: Exploring the Cosmic Frontier: Astrophysical Instruments for the 21st Century, ESO Astrophysical Symposia Series, ed by A.P. Lobanov (ESO, 2005), in press [arXiv:astro-ph/0606143] 45. A.P. Lobanov. In: 8th EVN Symp., Proc. of Sci., in press (2006) 46. C.B. Markwardt, J. Tueller, G.K. Skinner et al: ApJ 633, L77 (2005) 47. A.P. Marscher, F.E. Marshall, R.F. Mushotzky et al: ApJ 233, 498 (1979) 48. A.P. Marscher, S.G. Jorstad, J.L. Gómez et al: Nature 417, 625 (2002) 49. A.P. Marscher, S.G. Jorstad, M.F. Aller et al: AIP Conf. Ser. 714, 167 (2003) 50. A.P. Marscher: Mem.S.A.It. 76, 13 (2005) 51. J.R. Mattox, J. Schachter, L. Molnar et al: ApJ 481, 95 (1997) 52. J.R. Mattox, R.C. Hartman, O. Reimer: ApJS 135, 155 (2001) 53. J. Mcenery: ASP Conf. Ser. 350, 229 (2006) 54. K. Mitsuda, M. Bautz, H. Inoue et al: PASJ, in press (2006) 55. R.F. Mushotzky, C. Done, K.A. Pounds: ARA&A 31, 717 (1993) 56. G. Miley: ARA&A 18, 165 (1980) 57. F.B.S. Paerels, S.M. Kahn: ARA&A 41, 291 (2003) 58. A. Parmar. In: High-Resolution X-Ray Spectroscopy: towards Xeus and Con-X, ed by G. Branduardy-Raymont http://www.mssl.ucl.ac.uk/~gbr/workshop2/ 59. G. Piner: ASP Conf. Ser. 340, 55 (2005) 60. J.N. Reeves, A.C. Fabian, J. Kataoka et al: Astron. Nachr. 88, 789 (2006) 61. H. Röttgering: NAR 47, 405 (2003) 62. E. Ros, M. Kadler. In: Primer Encuentro de la Radioastronomı́a Española, ed by J.C. Guirado, I. Martı́-Vidal, J.M. Marcaide (Servicio de Publicaciones, Universidad de Valencia, Spain 2006) in press [arXiv:astro-ph/0608424] 63. R.T. Schilizzi. EAS Pub. Series 15, 445 (2005) 64. V. Semenov, S. Dyadechkin, B. Punsly: Science 354, 972 (2004) 65. D. Sowards-Emmerd, R.W. Romani, P.F. Michelson: ApJ 590, 109 (2003) 66. D. Sowards-Emmerd, R.W. Romani, P.F. Michelson, J.S. Ulvestad: ApJ 609, 564 (2004) 67. Y. Tanaka, H. Inoue, S.S. Holt: PASJ 46, L37 (1994) 68. Y. Tanaka, A. Fabian, H. Inoue et al: Nature 375, 679 (1995) 69. Y. Terzian, J. Lazio: Proc. SPIE 6267 (2006) 70. R.C. Vermeulen, E. Ros, K.I. Kellermann et al: A&A 401, 113 (2003) 71. K.A. Weaver, A.S. Wilson, C. Henkel, J.A. Braatz: ApJ 520, 130 (1999) 72. M.C. Weisskopf, H.D. Tananbaum, L.P. Van Speybroeck, S.L. O’Dell: Proc. SPIE 4012, 2 (2000) 73. A.R. Whitney, I.I. Shapiro, A.E.E. Rogers et al: Science 173, 225 (1971) 74. J.A. Zensus: ARA&A 35, 606 (1997) 75. J.A. Zensus, E. Ros, K.I. Kellermann et al: AJ 124, 662 (2002) 76. J.A. Zensus, E. Ros, M. Kadler et al: ASP Conf. Ser. 300, 27 (2003) Gaussian analysis of the CMB with the smooth tests of goodness of fit R.B. Barreiro1, J.A. Rubiño-Martı́n2, and E. Martı́nez-González1 1 2 Instituto de Fı́sica de Cantabria, CSIC – Universidad de Cantabria, Avda. de los Castros s/n, 39005, Santander, Spain, barreiro@ifca.unican.es, martinez@ifca.unican.es Instituto de Astrofı́sica de Canarias, C/ Vı́a Lactea s/n, 38200, La Laguna, Spain jalberto@iac.es Summary. The study of the Gaussianity of the cosmic microwave background (CMB) radiation is a key topic to understand the process of structure formation in the Universe. In this paper, we review a very useful tool to perform this type of analysis, the Rayner & Best smooth tests of goodness of fit. We describe how the method has been adapted for its application to imaging and interferometric observations of the CMB and comment on some recent and future applications of this technique to CMB data. 1 Introduction The study of the Gaussianity of the cosmic microwave background (CMB) fluctuations has become a very useful tool in constraining theories of structure formation. The standard inflationary scenario predicts Gaussian fluctuations whereas other competitive theories would imprint non-Gaussian signatures on the CMB (see [5] for a review). Therefore, the study of the Gaussianity of the CMB can help to discard or constrain some of these theories. Moreover, secondary effects (e.g. gravitational lensing, Rees-Sciama effect, Sunyaev-Zeldovich effect...), astrophysical emissions and systematics may as well leave non-Gaussian imprints on the CMB, which should not be confused with intrinsic non-Gaussianity. Given the importance of this type of analysis and taking into account that different methods may be sensitive to different kinds of non-Gaussianity, many tools have been developed for the study of the temperature distribution of the CMB. Among others, they include the Minkowski functionals [24], the bispectrum [18], wavelet techniques [4], geometrical estimators [27] or smooth tests of goodness of fit [3]. The interest for this type of analysis has increased even more since the release of the WMAP data [7]. A large number of different techniques have 178 R.B. Barreiro, J.A. Rubiño-Martı́n, and E. Martı́nez-González been applied to study whether these data follow or not a homogeneous and isotropic Gaussian random field, finding in some cases unexpected results. In particular, a significant number of works have reported deviations from Gaussianity and/or isotropy, whose origin is uncertain (e.g. [35, 17, 20, 12, 13, 15, 26, 37], see also [25] for a review). In this paper, we review the Rayner and Best smooth tests of goodness of fit for the study of the Gaussianity of the CMB. In section 2 we describe the test and how to adapt the method for its application to CMB observations. A discussion about current and future applications to different CMB datasets is given in section 3. Finally our conclusions are summarised in section 4. 2 The Rayner and Best smooth tests of goodness of fit Given a statistical variable X and n independent realizations xi , i = 1, ..., n, we want to test if X follows a given probability density function (pdf) f (x). The smooth tests of goodness of fit (gof) allows one to discriminate between a predetermined pdf f (x) (null hypothesis) and a second one that deviates smoothly from the former (alternative hypothesis). Among the possible forms for the alternative pdf, Rayner & Best [28, 29] consider: # " k X θi hi (x) f (x) , (1) fk (x, θ) = C(θ) exp i=1 where θ = (θ1 , ..., θk ) is a set of k parameters that allows for smooth deviations of the alternative hypothesis with respect to f (x), C(θ) is a normalisation constant that ensures that fk is normalised to 1 and hi form a complete set of orthonormal functions of f . Note that for θ = 0 we recover f (x), therefore, our statistical analysis consists on testing the null hypothesis H0 : {θ = 0} versus the alternative hypothesis H1 : {θ 6= 0}. To perform this analysis, the score statistic is used. This is a quantity which is closely related to the likelihood ratio (see e.g. [28]). For the Rayner & Best smooth tests of gof, the score statistic associated to the k alternative is given by Sk with Ui = = √1 n k X i=1 n X Ui2 (2) hi (xj ) (3) j=1 Large values of Sk (or of Ui2 ) reject the null hypothesis. In the case of testing if our data follow a Gaussian distribution of zero mean and unit dispersion, the hi are given by the (normalised) Hermite Chebishev Smooth tests of goodness of fit 179 polynomials.1 In this case, it is possible to write the Ui2 quantities in terms of the moments of order k, µk , of the data2 : U12 = nµ21 U22 = n (µ2 − 1)2 2 n (µ3 − 3µ1 )2 6 n U42 = (µ4 − 6µ2 + 3)2 24 U32 = (4) If the Gaussian hypothesis holds, the Ui2 follow a χ21 distribution when n → ∞. This allows one to determine easily the significance of any possible deviation from Gaussianity by comparing the value of the Ui2 of the data with a χ21 . We must point out that the proposed technique is designed to test if the data follow a univariate Gaussian. Thus, for optimality, it should be applied to independent data. However, the CMB signal is correlated at all scales and the noise may as well present correlations. Therefore, before applying the gof test, it is necessary to transform the data to make them as independent as possible. One possibility is to obtain the Cholesky decomposition of the correlation matrix of the data (including signal plus noise) C =PLLt and then multiply the xi by the inverse of the Cholesky matrix, i.e. yi = j L−1 ij xj . The constructed yi are uncorrelated, have zero mean and unit dispersion. Moreover if the data are Gaussian, they also follow a normal distribution and are independent. This decorrelation technique has been used for analysing the MAXIMA data with different smooth tests of gof [11, 1, 2]. Nevertheless, the preprocessing of the data has been improved in subsequent works through the use of a signal-to-noise decomposition, which is explained in the next subsection. 2.1 Signal-to-noise decomposition The signal to noise decomposition was introduced in the CMB field by [10], whereas [3] applied this formalism jointly with the gof test. This technique allows one to construct uncorrelated eigenmodes from the data which are also associated to a certain signal-to-noise ratio. Let us consider a set of CMB data di , i = 1, ..., n, where i corresponds to a given position in the sky. This can be written as di = si + ni (5) where si and ni are the contributions from the CMB signal and noise, respectively. The mean values of signal and noise are assumed to be zero and their correlation matrices are given by Sij = hsi sj i and Nij = hni nj i where the brackets indicate average over many realizations. The signal-to-noise eigenmodes are defined as 1 2 The form of the hi for other usual distributions (e.g. uniform, exponential) can be found in [28]. P k The moment of order k of the data is defined as µk = n j=1 yj /n 180 R.B. Barreiro, J.A. Rubiño-Martı́n, and E. Martı́nez-González t −1 ξ = RA LN d (6) where LN is the Cholesky matrix of N , i.e. N = LN LtN , and R is the rotation −t matrix that diagonalizes the matrix A = L−1 N SLN . The eigenvalues of this diagonalization are denoted by Ei . Let us now construct the quantities yi : ξi yi = √ 1 + Ei (7) It can be shown that these quantities are uncorrelated and have zero mean and unit dispersion. Moreover, if the data d are multinormal, then the yi are distributed according to a Gaussian pdf, since all the applied transformations are linear. In this case the yi are also independent. Therefore we are in the optimal conditions to apply the gof tests to the quantities yi . In addition, we also have information about the signal-to-noise ratio of the √ i eigenmode, which is given by Ei . This means that eigenmodes with low values of Ei are dominated by noise and may be discarded from the analysis. Therefore, in practice, the gof test will be applied to the subset of yi such that its signal-to-noise ratio is greater than a given threshold, i.e. Ei > Ecut . Thus, this decomposition allows us not only to obtain uncorrelated variables but also to select the fraction of the data where the signal contribution dominates over the noise. 2.2 Application to interferometer observations The previous technique has been adapted to deal with interferometric data by [3] and applied to VSA data in [30]. Let us consider an interferometer observing a small region of the sky at frequency ν, for which the flat-sky approximation is valid. In this case the complex visibility, which is the response of the interferometer at the considered frequency, is given by Z V (u, ν) = P (x̂, ν)B(x̂, ν) exp(i2πux̂)dx̂ (8) where x̂ corresponds to the angular position of the observed point on the sky and u is the baseline vector in units of the wavelength of the observed radiation. P (x̂, ν) is the primary beam of the antennas (normalized to unity at its peak) and B(x̂, ν) corresponds to the brightness distribution on the sky. Of course, for a realistic instrument, the effect of instrumental noise should be also taken into account. Therefore, the ith baseline ui of the interferometer will measure d(ui , ν) = V (ui , ν) + n(ui , ν) (9) where n(ui , ν) corresponds to the instrumental noise of the ui visibility. Let be N the total number of complex visibilities observed by the interferometer. Since the measured quantities are complex, the number of elements Smooth tests of goodness of fit 181 that constitute the data are Nd = 2N , corresponding to the real and imaginary parts of each observed visibility. Testing the Gaussianity of the measured visibilities is equivalent to testing the joint Gaussianity of their real and imaginary parts. Therefore the signal-tonoise decomposition can be applied directly to these quantities (so we will have a total of Nd eigenmodes). The correlation matrix S of the real and imaginary parts of V (ui , ν) (i.e. the correlation matrix of the signal) can be computed following the work of [21] whereas the noise correlation matrix is determined by the characteristics of the instrument. Once the signal-to-noise eigenmodes have been obtained, the gof technique can be applied to test the Gaussianity of these quantities (or of a subset of them with the highest signal-to-noise ratio). As in the previous case, if the data are distributed as a multinormal, the constructed eigenmodes are independent and follow a Gaussian distribution of zero mean and unit dispersion. A complementary analysis can also be performed on the phases of the decorrelated visibilities. If the data are Gaussian, the phases should follow a uniform distribution. This can be tested using the Rayner & Best smooth tests of gof by considering the appropriate hi in equation (2) (see [28, 3] for details). However, [3] found that, for their considered examples, the phase analysis was less sensitive to deviations from Gaussianity than the test based on the real and imaginary parts of the visibilities. 2.3 Some comments about the method One of the advantages of the Gaussianity analysis based on the gof test and the signal-to-noise formalism is that it is well suited for the study of many different kinds of CMB observations. In particular, it can be adapted to deal with most of the problematics found in real data. For instance, it is not affected by the presence of holes in the data or by the use of irregular masks and it can easily deal with anisotropic and/or correlated noise. Also, as already explained, it can be applied to imaging or interferometric data. Another interesting feature of the method is that it allows one to choose that fraction of the data with a signal-to-noise ratio above a certain threshold. In addition, as will be discussed in the next section, it is a very sensitive technique, being able to detect different type of deviations in the data (such as intrinsic nonGaussianity, systematic effects or anisotropy of the local power spectrum). The main shortcoming of the technique is the large amount of CPU required to calculate the signal-to-noise eigenmodes, since it involves the diagonalization of large matrices (of size n × n, where n is the number of data to be analysed). However, the method uses only a fraction of the eigenmodes (those whose signal-to-noise ratio is higher than a given threshold) and therefore it is not necessary to obtain all the eigenmodes and eigenvalues of the problem. To take advantage of this fact, [30] proposes the use of the Arnoldi algorithm which significantly speeds the calculation of the required yi . This method is based on the construction of a matrix H of dimension m × m (with m < n) 182 R.B. Barreiro, J.A. Rubiño-Martı́n, and E. Martı́nez-González such that it is possible to construct a good approximation to certain eigenvectors and eigenvalues of A from those of H. In particular, the eigenvectors that are well approximated correspond to those with higher eigenvalues. From these quantities it is also possible to construct those eigenmodes with higher signal-to-noise ratio, i.e., those that are kept for the analysis (see [30, 31] for details). This means that we have significantly reduced the computational cost of the analysis, since we are working with a matrix of size m × m instead of n × n. 3 Applications to CMB data The gof tests were firstly introduced in the CMB field by [11], which carried out a Gaussianity analysis of the MAXIMA data [19]. The results showed that the data were compatible with Gaussianity (see also [1, 2]). A more recent application of the Rayner & Best gof test has been carried out by [30], that present a Gaussianity analysis of the Very Small Array (VSA) data [34, 23, 16]. The VSA is an interferometer sited at the Teide Observatory (Tenerife) designed to observe the sky on scales going from 2◦ to 100 and operates at frequencies between 26 and 36 GHz (see [36] for a detailed description). In the analysis, most of the fields observed by the VSA were found to be compatible with Gaussianity. However, deviations from Gaussianity were detected in the U22 statistic in three cases. After a thorough analysis of the possible origins of these detections, the authors concluded that one of the deviations was associated to a residual systematic effect of a few visibility points, which, when corrected, have a negligible effect on the angular power spectrum. A second detection seemed to have its origin in a deviation of the local power spectrum of the considered field with respect to the power spectrum estimated from the complete dataset. This deviation was found at angular scales around the third angular peak (` = 700 − 900). If the affected visibilities were removed, a cosmological analysis based only on this modified power spectrum and the COBE data showed no differences except for the physical baryon density, which decreased by 10 per cent and got closer to the value obtained from Big Bang Nucleosynthesis. Finally, the third deviation from Gaussianity was found in observations of the Corona Borealis supercluster region [22]. In this case, the non-Gaussianity was identified as intrinsic to the data, probably due, at least in part, to the presence of Sunyaev-Zeldovich emission in the region. This result has been later confirmed with the measurements of the MITO telescope in this region [6]. A combined maximum likelihood analysis of the MITO and the VSA data provided a weak detection of a faint signal compatible with a SZ effect, characterized by a Comptonization parameter of −6 y = (7.8+5.3 , at 68% CL. −4.4 ) × 10 An application of the gof technique to the Archeops data is currently ongoing [14]. Archeops is a balloon-borne experiment, which is dedicated to Smooth tests of goodness of fit 183 measure the CMB temperature anisotropies from large to small angular scales [8, 9]. It has also been designed as a test bed for the forthcoming Planck high frequency instrument. The preliminary results show the good performance of the method, that is able to deal with the presence of anisotropic and correlated noise in the data. The application of the gof technique to the WMAP data [7] is of great interest and is currently in progress. Due to the large amount of data observed by this experiment, a whole sky analysis at full resolution is unfeasible, due to the large computational resources required for the signal-to-noise decomposition. However, two types of complementary tests are possible: an analysis of the full-sky at low-resolution and a study of small regions of the sky at high resolution. Given the sensitivity of the gof tests to detect deviations from a homogeneous and isotropic Gaussian random field, this analysis could shed new light on some of the anomalies reported for the WMAP data. 4 Conclusions We have reviewed the Rayner & Best smooth tests of goodness of fit and its applications to CMB data. One of the most interesting features of this method is that it can deal with most of the problematics found in real data such as the use of irregular masks or the presence of anisotropic and/or correlated noise. In addition, it has been adapted to deal either with imaging or interferometric observations. The main shortcoming of the technique is the large computational cost required to perform the signal-to-noise decomposition of the data. However, this problem can be significantly alleviated by the use of approximate methods such as the Arnoldi algorithm. The recent and current applications of the gof tests to different datasets are showing its good performance. Most notably, the method has been able to detect deviations from a homogeneous and isotropic Gaussian field in the VSA data, which were associated to very different origins: residual systematics, a deviation of the local power spectrum with respect to the global one and non-Gaussianity intrinsic to the data. It is important to mention that Gaussianity analyses had already been performed in the VSA dataset using other methods [32, 33] but neither the residual systematics nor this small deviation of the power spectrum were detected. Therefore we believe that this method constitutes a very useful tool for the statistical analysis of CMB data. Acknowledgements: RBB and EMG thank financial support from the Spanish MEC project ESP2004-07067-C03-01. References 1. A.M. Aliaga, E. Martı́nez-González, L. Cayón, F. Argüeso, J.L. Sanz, R.B. Barreiro: New Ast. Rev., 47, 821 (2003) 184 R.B. Barreiro, J.A. Rubiño-Martı́n, and E. Martı́nez-González 2. A.M. Aliaga et al.: New Ast. Rev., 47, 907 (2003) 3. A.M. Aliaga, J.A. Rubiño-Martı́n, E. Martı́nez-González, R.B. Barreiro, J.L. Sanz: MNRAS, 356, 1559 (2005) 4. R.B. Barreiro, M.P. Hobson: MNRAS, 327, 813 (2002) 5. N. Bartolo, E. Komatsu, S. Matarrese, A. Riotto: Phys. Rep., 402, 103 (2004) 6. E.S. Battistelli et al.: ApJ, 645, 826 (2006) 7. C.L. Bennett et al.: ApJS, 148, 1 (2003) 8. A. Benoı̂t et al.: Astropart. Phys., 17, 101 (2002) 9. A. Benoı̂t et al.: A&A, 399, L19 (2003) 10. J.R. Bond: Phys. Rev. Lett., 74, 4369 (1995) 11. L. Cayón, F. Argüeso, E. Martı́nez-González, J.L. Sanz: MNRAS, 344, 917 (2003) 12. L. Cayón, J. Jin, A. Treaster: MNRAS, 362, 826 (2005) 13. C.J. Copi, D. Huterer, D.J. Schwarz, G.D. Starkman: MNRAS, 367, 79 (2006) 14. A. Curto, J. Aumont, J.F. Macı́as-Pérez, E. Martı́nez-González, R.B. Barreiro, D. Santos: in preparation 15. M. Cruz, M. Tucci, E. Martı́nez-González, P. Vielva: MNRAS, 369, 57 (2006) 16. C. Dickinson et al.: MNRAS, 353, 732 (2004) 17. H.K. Eriksen, F.K. Hansen, A.J. Banday, K.M. Górski, P.B. Lilje: ApJ, 605, 14 (2004) 18. P.G. Ferreira, J. Magueijo, K.M. Górski: ApJ, 503, L1 (1998) 19. S. Hanany et al.: ApJ, 545, L5 (2000) 20. F.K. Hansen, A.J. Banday, K.M. Górski: MNRAS, 354, 641 (2004) 21. M.P. Hobson, K. Maisinger: MNRAS, 334, 569 (2002) 22. R. Génova-Santos et al.: MNRAS, 363, 79 (2005) 23. K. Grainge et al.: MNRAS, 341, L23 (2003) 24. E. Komatsu et al.: ApJS, 148, 119 (2003) 25. E. Martı́nez-González: Cosmic Microwave Background anisotropies: the power spectrum and beyond. In: Data Analysis in Cosmology, ed by V. Martı́nez, E. Martı́nez-González, M.J. Pons-Borderı́a, E. Saar (Springer-Verlag), in press 26. J.D. McEwen, M.P. Hobson, A.N. Lasenby, D.J. Mortlock: MNRAS, 371, L50 (2006) 27. C. Monteserı́n, R.B. Barreiro, E. Martı́nez-González, J.L. Sanz: MNRAS, 371, 312 (2006) 28. J.C.W. Rayner, D.J. Best: Smooth Tests of Goodness of Fit (Oxford University Press, New York, 1989) 29. J.C.W. Rayner, D. J. Best: International Statistical Rev., 58, 9 (1990) 30. J.A. Rubiño-Martı́n et al.: MNRAS, 369, 909 (2006) 31. Y. Saad: Numerical Methods for Large Eigenvalue Problems (Manchester Univ. Press, Manchester, 1992) 32. R. Savage et al.: MNRAS, 349, 973 (2004) 33. S. Smith et al.: MNRAS, 352, 887 (2004) 34. A.C. Taylor et al.: MNRAS, 341, 1066 (2003) 35. P. Vielva, E. Martı́nez-González, R.B. Barreiro, J.L. Sanz, L. Cayón: ApJ, 609, 22 (2004) 36. R.A. Watson et al: MNRAS, 341, 1057 (2003) 37. Y. Wiaux, P. Vielva, E. Martı́nez-González, P. Vandergheynst: Phys. Rev. Lett. 96, 151303 (2006) Dark matter in galaxy clusters N. Benı́tez Instituto de Astrofı́sica de Andalucı́a(CSIC), C/Camino Bajo de Huétor 50, Granada 18008, Spain benitez@iaa.es Summary. Evidence for dark matter in galaxy clusters was first discovered by Zwicky more than 70 years ago. Since then galaxy clusters have often proved to be crucial laboratories to study the properties and nature of this mysterious component of the universe. The existence of dark matter in the core of galaxy clusters in solidly established, and it is not even challenged by alternative theories of gravity like MOND. It is also becoming increasingly clear that this cluster dark matter is collisionless, and that the current CDM paradigm nicely fits the existing observations. To mount a successful challenge against the standard CDM model, MOND advocates will have to find a feasible explanation for the cluster dark matter (neutrinos as dark matter are being resurrected for this purpose) and also develop the machinery to make clear, unambiguous predictions about the dynamical and lensing properties of galaxy clusters. 1 Some history about dark matter The history of our evolving knowledge about dark matter is an interesting one, and detailed accounts can be found in [26] and [35]. [16] was the first to estimate the mass of galaxies based on the observed motions of the stars and gas and the energy needed to gravitationally bind them. His observations were obviously confined to the bright, central of the galaxies and therefore missed any mass contained in the fainter outskirts. In 1933 Zwicky calculated the dispersion of the radial velocities of 8 galaxies in the Coma cluster and found the unexpectedly large value of σ = 1019 ± 360km/s. He concluded, using Hubble’s results for galaxy masses and the “best” value of the Hubble parameter at the time (H0 = 558kms−1 Mpc−1 ), that the total mass contained in the Coma cluster had to be 400 times larger than the mass contained in the bright parts of individual galaxies.With the WMAP value of H0 = 70kms−1 Mpc−1 , the number inferred by Zwicky would be ≈ 50. Three years later, [32] found the same phenomenon in the Virgo cluster. Zwicky did not advance any hypothesis about the origin of this “dunkle materie”but Smith speculated that it “represents a great mass of internebular material 186 N. Benı́tez within the cluster”. In 1939, and without being aware of the results of [37], [3] discovered that the outer regions of Andromeda were rotating much faster than expected, which he interpreted as either a very large mass to luminosity ratio or dust absorption. [21] studied the rotation and surface brightness of the edge-on S0 galaxy NGC3115 and again found very high mass-to-light ratios in the external parts of the galaxy. These findings went practically unnoticed and dark matter had to wait another generation to be rediscovered by [17], who used the dynamics of M31 and the Galaxy to infer that the mass of the Local Group had to be 3−4 times larger that the combined mass of the two dominant galaxies. They proposed that the “missing mass” as it was called at the time, was in the form of hot, 5 × 105 K gas. [25] and [33] studied the statistical distribution of separations and velocity differences of galaxy pairs, and again saw evidence of considerable amounts of dark matter associated with them. In the early 60’s [1] proposed that galaxy clusters were unstable configurations with positive energy in the process of forming new galaxies and which would eventually disperse. However, the presence of a very large fraction of old, elliptical galaxies in the clusters made this hypothesis untenable ([34]). Another decade went by until [29] measured the radial velocities of HII regions in Andromeda to find that its rotation velocity rose very fast, up to 225kms−1 at 400pc. Although these results were met with skepticism, the evidence from different sources was adding up. [22] showed that a flat, selfgravitating disk with the rotation curve of a spiral galaxy is dynamically unstable and that most of the mass in the inner parts of a spiral galaxy has to be in a component more stable and with higher velocity dispersion that the observed stars. It was natural to extrapolate this idea of an inner dark matter halo to match the rotation curve results in the outer regions of galaxies. Ostriker, [23] noted that the mass-to-light ratios of galaxies grow with increasing radius, and the total mass contained in this objects is large enough to have cosmological significance. Finally, [28] extended the rotation curve measurement in Andromeda to 30kpc using 21 − cm line observations, and their results, together with additional work by Rubin & Ford clearly showed that the rotation curve of Andromeda did not present the drop-off in circular speed expected from Newtonian mechanics, but remained constant for radii of 16 − 30kpc, again confirming that the outer regions of the galaxy contained enormous amounts of mass not in the form of light-emitting gas and stars. In addition, [11] noted that the X-ray emitting gas in rich galaxy clusters was not enough to bind these objects. Therefore by 1975 most astronomers had to confront the fact that if the Newton-Einstein theory of gravitation was correct, dark matter was necessary to explain the dynamical behavior of galaxies, groups and clusters. The most natural hypothesis was that this “missing mass” was in the form of baryons: diffuse gas, late M dwarfs, brown dwarfs or compact objects like white dwarfs, neutron stars or even black holes. However, it was soon clear Dark matter in galaxy clusters 187 that the stringent constraints imposed on the baryonic content of the universe by the big bang nucleosynthesis ([14]) excluded this possibility. [13] first proposed neutrinos as candidate for dark matter, and as [9] found, they could in principle account for the observed properties of galaxy clusters. There was considerable interest in this hypothesis in the late 70’s and early 80’s but it soon fell out of favor. The new experimental limits on the mass of the electron-dominated family implied that they did not represent an important cosmological contribution, there were considerable problems to form galaxies in a neutrino-dominated universe and it became increasingly difficult to explain the properties of dark matter in dwarf galaxies using neutrinos. But perhaps the most important factor in the demise of hot dark matter, as neutrinos were called, was an almost text-book application of Kuhn’s paradigm shifts: the arrival of the cold dark matter hypothesis. 2 Cold dark matter (CDM) It should be noted that at the time the CDM paradigm was born, the main concern of cosmologists was not to explain the dynamical behavior of cosmic structures, but to fill the universe up to the critical density Ω = 1 predicted by inflation. A review by [24] mentions that the desired properties of dark matter include very weak interaction with normal matter and the fact that it does not show up in our dynamical estimates requires that it be less clustered than our usual dynamical probes, i.e. galaxies. That is, theoreticians were primarily looking for a candidate which would close the universe. It was soon clear that baryons were not enough for this purpose, and when the same happened with neutrinos, they fell off the radar screen specially as GUTs and other theories started to provide possible dark matter candidates. CDM is formed, by definition, by particles which were very slow at the time when galaxy formation started. Many types of cold dark matter have been proposed by theoretical physicists, although none of them have been detected directly. However, the cold dark matter hypothesis elegantly explains most of the observations mentioned in the previous section and its problems are minor compared with those of other alternatives so we know speak of the “standard CDM cosmology”. CDM nicely explains the formation of cosmic structures starting from adiabatic fluctuations, and provide satisfactory fits to their individual mass distributions. [18] showed that CDM haloes display a so-called universal profile with the following shape: ρ(r) = ρs (r/rs )(1 + r/rs )2 Minor modifications around this shape have been proposed,specially for the inner slope, but this profile represents well the expected radial density of CDM haloes. 188 N. Benı́tez 3 Modified Newtonian Dynamics (MOND) As we mentioned before, dark matter was introduced to explain the dynamics of galaxies and clusters within the framework of Newtonian-Einstein gravity. However a different possibility is changing the laws of gravity. Most physicists have a strong prejudice against new physics, and for good reason, since the vast majority of new theories does not survive first contact with observations or experiments. In the middle of the debate between CDM and HDM, [20] proposed to modify gravity so that the acceleration experimented by a test particle around a mass M is given by a2 /a0 = GM r−2 for a < a0 instead of the customary a = GM r−2 , which still holds for high accelerations a > a0 . This ad hoc change, intended to fit the rotation curves of bright galaxies, has been extremely successful in making predictions for dwarf galaxies. It should be stressed that if, in a mental experiment, one forgets about all other cosmological evidence and focuses only on the dynamical behavior of galaxies, it seems clear that MOND is a more successful explanation that the standard CDM picture (see [19] for a review), and any successful DM theory would have to explain why galaxy dynamics are so well described by MOND far beyond its initial range of application. However there are many other lines of evidence in which MOND does not fare so well. A relativistic version of MOND, an essential requisite for the theory to be considered seriously, has been lacking until recently ([4]) and it is still being explored. It is not straightforward to use this theory to calculate cosmological predictions, and therefore it is far from clear whether it can match the overall cosmological picture with the degree of success reached by standard CDM. But perhaps the biggest conceptual hurdle for MOND is that it cannot fully explain the dynamical and lensing properties of massive galaxy clusters without introducing some kind of dark matter ([31], [27]). This fact does not “disprove” MOND, it is perfectly possible to have dark matter and modified gravity coexisting, but Occam’s razor, a very good guide for similar dilemmas, tells us that if we have to introduce dark matter anyway, there is little reason to add more hypothesis and tinker with the laws of physics. 4 CDM and lensing clusters As we have explained above, galaxy clusters are privileged laboratories for testing CDM and its alternatives. Strong lensing, which produces multiple images of the same source, is observed when the projected mass of a cosmic structure approximately exceeds 1.0g/cm−2. Most rich galaxy clusters display densities above this limit, and giant arcs and multiple images are common around them. The Abell cluster A1689 has the largest known Einstein radius, approximately 5000 , and it was one of the first targets to be observed in 2002 by the Advanced Camera for Surveys ([12]) aboard the Hubble Space Telescope. The observations are much deeper, have a wider field, better image quality Dark matter in galaxy clusters 189 and richer color information that previous data obtained with WFPC2 (see Fig. 1). [6] identified 30 different multiply lensed background galaxies which yielded 106 different images. The galaxies span a very wide redshift range, 1.0 < z < 5.5 and clearly delineate the radial and tangential critical curves. This excellent data set provides the most accurate measurement of the central mass profile of any cluster and can be very well fit by a NFW profile although with a relatively high concentration, cvir ≈ 8 ± 2 ([7]). This dataset has been studied by several authors ([10, 36, 30, 15], using independent approaches which have yielded similar results. It is of course dangerous to extract strong conclusions from such an unusual object as A1689, and there remains the problem of the concentration index, considerable larger than expected for this kind of massive haloes but it is clear that there is a more than reasonable agreement between CDM simulations, as represented by the NFW profile, and the observations of strong lensing in Abell 1689. Recently [8] and [5] have published lensing analysis of the merging cluster 1E0657-558 (z = 0.296). Their results illustrate how most of the mass in the cluster is concentrated around the position of the two clumps of cluster galaxies, and does not follow the distribution of the X-ray emitting gas. Their results add strong evidence, in a very visual way, to the picture already inferred from previous observations and theoretical considerations: a) there is dark matter in galaxy clusters and b)it is collisionless. However, it does not really change the perception we have of MOND, since as we have seen previously, it was already well known that MOND requires dark matter in the central parts of galaxy clusters. A recent paper, [2], even argues that in systems lacking spherical symmetry, the predictions of the relativistic version of MOND, TeVeS can mimic observations like those of 1E0657-558. 5 Conclusions The existence of dark matter in the core of galaxy clusters in solidly established, and it is not even challenged by alternative theories of gravity like MOND. It is becoming increasingly clear that this cluster dark matter is collisionless, and the current dark matter paradigm nicely fits the existing observations. If the proponents of MOND or other alternative gravity theories want to mount a successful challenge against the standard CDM model, they will have to find a feasible explanation for this dark matter (neutrinos as dark matter are being resurrected for this purpose) and also develop the machinery to make clear, unambiguous predictions about the dynamical and lensing properties of galaxy clusters. But CDM advocates cannot rest on their laurels yet: the intriguing ability of MOND to precisely fit the rotation curves of galaxies using their light profiles has to be explained and understood within the CDM paradigm. 190 N. Benı́tez Fig. 1. Color image obtained by combining the 4 filter ACS observations of Abell 1689 Dark matter in galaxy clusters 191 References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. Ambartsumian, V.A., The Astronomical Journal, 66, 536 (1961) Angus, G. W., Famaey, B., & Zhao, H. S. 2006, MNRAS, 371, 138 Babcock, H.W., Lick Obs. Bull, 19, 41 (1939) Bekenstein, J. D., PhysRevD, 70, 083509 (2004) Bradac, M., et al., ArXiv Astrophysics e-prints, arXiv:astro-ph/0608408 (2006) Broadhurst, T., et al. The Astrophysical Journal, 621, 53 (2005) Broadhurst, T., Takada, M., Umetsu, K., Kong, X., Arimoto, N., Chiba, M., & Futamase, T. 2005, ApJL, 619, L143 (2005) Clowe, D., Bradač, M., Gonzalez, A. H., Markevitch, M., Randall, S. W., Jones, C., & Zaritsky, D., ApJL, 648, L109 (2006) Cowsik, R. & MacClelland, J., The Astrophysical Journal, 180, 7 (1973) Diego, J. M., Sandvik, H. B., Protopapas, P., Tegmark, M., Benı́tez, N., & Broadhurst, T., MNRAS, 362, 1247 (2005) Einasto, J., Kaasik, A. & Saar, J., Nature, 250, 309 (1974) Ford, H. C., et al., Proceedings of the SPIE, 4854, 81 (2003) Gershtein, S.S. & Zel’dovich,Ya.B. 1966, JETP Lett. 4, 174 (1966) Gott, J. R., III, Schramm, D. N., Tinsley, B. M., & Gunn, J. E., The Astrophysical Journal, 194, 543 (1974) Halkola, A., Seitz, S., & Pannella, M. 2006, MNRAS, 372, 1425 (2006) Hubble, E. P., The Astrophysical Journal, 64, 321 (1926) Kahn, F.D. & Woltjer, L. The Astrophysical Journal, 130, 705 (1959) Navarro, J. F., Frenk, C. S., & White, S. D. M. The Astrophysical Journal, 490, 493 (1997) McGaugh, S., ArXiv Astrophysics e-prints, arXiv:astro-ph/0606351 (2006) Milgrom, M. The Astrophysical Journal, 270, 365 (1983) Oort, J.H., The Astrophysical Journal, 91, 273 (1940) Ostriker, J.P. & Peebles, P.J.E., The Astrophysical Journal, 186, 467 (1973) Ostriker, J.P., Peebles, P.J.E. & Yahil, A., The Astrophysical Journal, 193, L1 (1974) Ostriker, J. P., ARA&A, 31, 689 (1993) Page, T., The Astrophysical Journal, 132, 910 (1960) Peebles, P. J. E., “Principles of Physical Cosmology”, Princeton Series in Physics, Princeton, NJ: Princeton University Press, (1993) Pointecouteau, E., & Silk, J. MNRAS, 364, 654 (2005) Roberts, M.S. & Whitehurst, R.N., The Astrophysical Journal, 201, 327 (1975) Rubin V.C. & Ford, W.K., The Astrophysical Journal, 159, 379 (1970) Saha, P., Read, J. I., & Williams, L. L. R. 2006, ApJL, 652, L5 (2006) Sanders, R. H. MNRAS, 342, 901 (2003) Smith, S. The Astrophysical Journal, 83, 23 (1936) van den Bergh, S., The Astronomical Journal, 66, 566 (1961) van den Bergh, S., Zs. f. Astrophys., 55, 21 (1962) van den Bergh, S., PASP, 111, 657 (1999) Zekser, K. C., et al., The Astrophysical Journal, 640, 639 (2006) Zwicky, F. Helv. Phys. Acta, 6, 110 (1933) Cosmology with the Largest Scale Structures: Probing Dark Energy F. J. Castander and the Dark Energy Survey Collaboration Institut de Ciències de l’Espai (CSIC-IEEC), Campus UAB, Facultat de Ciències, 08193 Bellaterra, Barcelona, Spain, fjc@ieec.fcr.es Summary. The understanding of Dark Energy is one the the great challenges of Cosmology and can only be phenomenologically studied sampling the largest scales structures of the universe. The four more promising techniques to study dark energy are the study of clusters of galaxies, galaxy clustering and particularly baryon acoustic oscillations, gravitational lensing and supernovae. The Dark Energy Survey will perform a very large and deep photometric survey of 5000 deg2 in four bands (g, r, i, z), with the aim of producing the first precise characterization of the properties of dark energy using these four complementary techniques. For this purpose it will build a new 3 deg2 field of view camera to be placed on the Blanco 4m telescope prime focus. It will use new thick CCDs that are very sensitive in the red allowing an efficient survey in the z band and thus probing to large redshifts, z ∼ 1. 1 Introduction Observational Cosmology has made great advances in recent years. The combination of several observational probes now provides a coherent picture of the universe we live in and allow us to measure cosmological parameters with high level of precision. In particular, we now believe that 75% of the energymatter content of the Universe is of some weird form unknown to us (called dark energy) that causes the Universe to accelerate. Moreover, another 20% is some kind of unknown matter detectable by its gravitational interaction which we call dark matter. We are faced with the embarrassment of ignoring what constitutes 95% of the Universe. Understanding these mysterious components is one of the great challenges of Cosmology and Theoretical Physics. Within this context in December 2003, the National Optical American Observatory (NOAO) released an announcement of opportunity to build in partnership with NOAO a major community instrument for the Blanco 4m telescope at Cerro Tololo (CTIO), Chile. In exchange, 30% of the Blanco 4m time for 5 years would be awarded to the instrument providers. In response to 194 F. J. Castander and the Dark Energy Survey Collaboration this announcement the Dark Energy Survey 1 (DES) collaboration was formed and proposed the construction of the Dark Energy Camera (DECam) a 3 deg2 field of view camera to be placed at the CTIO 4m. The DES proposal was selected. The DES collaboration has grown since then and now is composed of 14 institutions and 81 participants. 2 Probing Dark Energy Although we do not know what dark energy is (e.g., cosmological constant, quintessence, phantom energy or modifications of General Relativity), we can measure its observational consequences. It is now customary to parameterize the effect of dark energy by its effective equation of state that appears in the Friedman equation, w = P/ρ. We can probe dark energy through its dependence on the expansion history of the universe H(z) and the growth factor of structure g(z). We can measure H(z) through its integrals: the comoving, angular-diameter and luminosity distances and the volume element. Experiments that measure these quantities are considered geometrical probes and include supernovae, clusters of galaxies, baryon acoustic oscillations and weak lensing. Experiments that measure the growth rate of structure include weak lensing and clusters of galaxies. A combination of experiments that use geometrical and growth rate techniques is desirable in order to understand the nature of dark energy (whether it is a strange fluid or a modification of gravity). There are thus four main complementary techniques to probe dark energy: Galaxy Clusters: The formation and evolution of structure is seeded by initial perturbations and driven by gravitational instability in a dynamically evolving universe. Massive structures observed in the universe bear the marks of these three influences: initial perturbations, the process of gravitational collapse and the evolving underlying metric. Clusters of galaxies form from the high mass end tail of the density perturbations and therefore their abundance and evolution are very sensitive to cosmological parameters. For a given sample under study the cosmological sensitivity comes from three basic elements: the volume sampled, the abundance evolution and the selection function. Clusters are mainly formed by dark matter, hot gas and galaxies. Surveys select clusters using some observable of these components: luminosity, number of galaxies, SZ flux, weak lensing shear,... None of these observables measures directly the cluster mass (even weak lensing does not, as it measures projected mass), which is the cluster property necessary to obtain cosmological information. A detail understanding of the correlation of the observed property with mass and its scatter is thus necessary for cosmological tests. Moreover, in a flux-limited survey the selection function also depends on the luminosity distance which also depends on cosmology. One possible way to overcome the 1 http://www.darkenergysurvey.org/ Cosmology with the Largest Scale Structures: Probing Dark Energy 195 uncertainties of the observed property and mass in constraining cosmological parameters is the use of ”self-calibrating” techniques (e.g., [9, 12]), in which the redshift distribution and the cluster power spectrum can be used to overcome the exact knowledge of the observable-mass relation and its evolution. Weak Gravitational Lensing: The bending of light by foreground mass concentrations shears the images of distant source galaxies. Large scale structures generate correlated shear that can be studied measuring shear-shear correlations (also known as cosmic shear). Since the foreground dark matter is associated to large degree with foreground galaxies, one can also measure the angular correlation between foreground galaxies and source galaxy shear (galaxy-shear correlations or galaxy-galaxy lensing). These weak lensing techniques provide powerful probes of dark energy as they are sensitive to the cosmic expansion history through both the geometry of the universe and the growth rate of structure [10, 11]. Galaxy Clustering: In the linear regime, we can write the galaxy power spectrum as a function of the initial dark matter power spectrum from the early Universe, the scale-dependent transfer function for dark matter perturbations, the scale-independent linear perturbation growth function and the bias. Through these dependences the galaxy clustering provides direct information about cosmology. The transfer function give us information on the nature and amount of dark matter, including the neutrino mass. Moreover, characteristic scales in the galaxy power spectrum, including the sound horizon scale at matter-radiation equality and the baryon acoustic oscillations, provide physically standard rods that can be calibrated and used to measure the angular diameter distance as a function of redshift [4, 5]. In addition, cross-correlation of Cosmic Microwave Background (CMB) data sets with galaxies as tracers of potential wells probes dark energy through the integrated Sachs-Wolfe (ISW) effect [6]. Supernovae: The study of supernova (SN) light curves to measure the expansion history of the universe has rapidly become a standard in cosmological studies. Studies of nearby SNe [7] provided the basis for development of methods of using Type Ia SNe as precision distance indicators based on their maximum apparent magnitudes and decline rates or “stretch” factors [7, 15, 13], and the application of these methods to studies of high redshift SNe provided the first direct evidence of the accelerating expansion of the Universe [16, 14]. Much of the power in determination of cosmological parameters with SNe relies upon the complementary nature of the confidence contours derived from SN studies with those derived from studies of galaxies and large scale structure and CMB data. 3 The Dark Energy Survey All four techniques described above require sampling large volumes (area & depth). In order to measure the evolution of the expansion rate and the growth 196 F. J. Castander and the Dark Energy Survey Collaboration of structure one needs to sample different redshifts, which requires wide area for low z and depth for high z. Clusters of galaxies are rare objects and large volumes at each redshift are needed to measure their abundance evolution. Baryon acoustic oscillations (BAO) are imprinted on scales of ∼100 h−1 Mpc. Weak lensing requires a sufficient number of background galaxies with well measured shapes. SNe searches need area at low z and depth at high z. For weak lensing good image quality is also necessary and repeat imaging for SNe. These techniques also require accurate redshift determination of the galaxies sampled. Unfortunately, the large number of objects needed makes it impossible to perform a spectroscopic survey with current instrumentation to probe dark energy. The best alternative is photometric surveys with good spectral coverage and good photometry as to deliver accurate photometric redshifts. The Dark Energy Survey (DES) main goal is to characterize dark energy measuring its equation of state through the four independent probes mentioned above carrying out a photometric survey. The DES has been defined to have the following characteristics: The DES will cover 5000 deg2 of the Southern Hemisphere divided in three subareas (see Figure 1a). The SPT overlapping area will be 4000 deg2 , the SDSS equatorial stripe 82 will cover 200 deg2 and a connecting area of 800 deg2 . The survey is to be completed in 5 years using 30% of the Blanco time. To carry out the SN program, there will be a 40 deg2 area sampled every three nights in one filter and every six nights in two filters. The survey will be done in 4 bandpasses, the SDSS griz The limiting magnitudes will be g ∼ 24.6, r ∼ 24.1, i ∼ 24.3 and z ∼ 23.9, for small galaxies at 10σ. Observations will be taken in 525 nights during the September-February period, starting in 2010 and finishing in 2015. The data will be public with an expected release schedule similar to the SDSS, e.g. one year. 3.1 Photometric Redshifts In order to achieve its scientific goals, the Dark Energy Survey will need to obtain accurate galaxy photometric redshifts (photo-z’s) derived from the DES griz imaging data. Detailed understanding of the photo-z error distributions as functions of galaxy magnitude, redshift and type, will be important for obtaining accurate cosmological parameter constraints. In particular, very large completed and ongoing spectroscopic redshift surveys will be available before the DES observations, and they will provide the data sets needed for accurate calibration and measurement of photo-z’s and photo-z errors, down to the DES photometric limit. Cosmology with the Largest Scale Structures: Probing Dark Energy 197 3.2 The South Pole Telescope One important point of the DES is its synergy with the South Pole Telescope 2 (SPT) survey. The selection of the sample is of paramount importance to study dark energy using clusters. The DES will select clusters using photometric techniques. Complementary to this selection the DES will also study a Sunyaev-Zeldovich (SZ) selected sample detected with the South Pole Telescope [17]. The SPT is a radio telescope to be installed in the South Pole that will be devoted to the detection of galaxy clusters through the SZ effect. The SZ effect is a change in the CMB radiation spectrum when the CMB photons pass through the hot intracluster medium and are inverse Compton scattered by the hot electrons producing a distinct spectral distribution in the CMB. The SPT will detect vast amounts of massive galaxy clusters (> 1014 M ) with a well defined selection function. Fig. 1. a) left: footprint of the DES survey area over-plotted on a dust extinction map. b) right: cross section of the DECam 4 The Dark Energy Camera Figure 1b shows a cross section of DECam. The major components of the instrument are a 500 megapixel optical CCD camera, a wide-field optical corrector (2.2o diameter field of view), a 4-band filter system with SDSS g, r, i, and z filters, guide and focus sensors mounted on the focal plane, low noise CCD readout, a cryogenic cooling system to maintain the focal plane at 180 K as well as a data acquisition and instrument control system to connect to the Blanco observatory infrastructure. The camera focal plane will consist of sixty-two 2K x 4K CCDs (0.27”/pixel) arranged in a hexagon covering an imaging area of 3 deg2 . Smaller format CCDs for guiding and focusing will 2 http://spt.uchicago.edu/ 198 F. J. Castander and the Dark Energy Survey Collaboration be located at the edges of the focal plane. To efficiently obtain the z-band images at high redshift (z ∼ 1) we selected the fully depleted, high resistivity, 250 micron thick silicon devices that have been designed and developed at the Lawrence Berkeley National Laboratory (LBNL) [8]. The QE of these devices is > 80% QE in the z band, roughly a factor of 10 higher than traditional thinned astronomical devices. The optical corrector design consists of five fused silica lenses that produce an unvignetted 2.2 deg diameter image area, which is calculated to contribute < 0.4” FWHM to the point-spread function. DECam will be installed in a new prime focus cage. We have adopted the Monsoon CCD readout system developed by NOAO that is being tested and adapted to DECam at FNAL and Spain. 5 The VISTA Dark Energy Survey (VDES) In January 2006, ESO released a call for proposal for public near-infrared surveys to be carried out at the Visual and Infrared Survey Telescope for Astronomy (VISTA). VISTA is a 4m telescope that is being built at Paranal with a near infrared camera of 16 2k × 2k detector with pixel scale of 0.34”/pixel giving a field of view of non-contiguous 0.6 deg2 that tile into a rectangular 1.6 deg2 . We submitted a VISTA Public Survey proposal to ESO to image the DES area in the near infrared, with the science goals of studying dark energy and providing a legacy survey for the community. Imaging the DES area in the near infrared adds valuable information about the SED of the galaxies that improve the photometric redshifts determinations, especially at z > 1, thus improving the constraints that DES can place on dark energy. The ESO Public Survey Panel recommended to accommodate the VDES proposal into the VISTA Hemisphere Survey (VHS, PI: Richard McMahon). The final VHS proposal was submitted at the end of September 2006 including near-infrared imaging of the DES area. 6 The Dark Energy Task Force Report The Dark Energy Task Force (DETF) was appointed by AAAC and HEPAP as joint subcommittee to advise three agencies: DOE, NASA and NSF. The DETF requested white papers for dark energy experiments from the community. They received 50 white papers. The Dark Energy Survey submitted an experimental white paper [1] in which the forecasts for DE constraints are estimated (see Table 1) and two theoretical white papers [2, 3] In their report, they “strongly recommend ... an aggressive program to explore dark energy”. They considered four techniques, those of the DES, and defined four stages of projects: Stage I = completed, Stage II = on-going, Stage III = near term, mid cost, proposed and Stage IV = LST, JDEM, SKA. They “recommend that the ... program have multiple techniques at every stage” Cosmology with the Largest Scale Structures: Probing Dark Energy 199 Table 1. Example forecast marginalized 68% CL statistical DES constriants on constant w Method/Prior Uniform WMAP Plank Clusters abundance 0.13 0.10 0.04 w/ WL mass calibration 0.09 0.08 0.02 Weak Lensing shear-shear (S-S) 0.15 0.05 0.04 galaxy-shear (G-S)+G-G 0.08 0.05 0.03 S-S+G-S+G-G 0.03 0.03 0.02 S-S+bispectrum 0.07 0.03 0.03 Galaxy angular clustering 0.36 0.20 0.11 Supernovae Ia 0.34 0.15 0.04 and that stage III should start immediately with projects similar to DES and WFMOS. 7 Simulations We are producing a series of catalog and image simulations for the Dark Energy Survey, for the purposes of helping us to develop our science analysis codes and data reduction pipelines prior to the start of survey. These simulations will also serve to help us characterize both the statistical and systematic errors inherent in the cosmological parameter analysis techniques for each of the 4 DES key projects. We plan to have yearly cycles of new catalog and image simulations, followed by science analysis and data reduction challenges carried out using the simulation outputs. The level of scale and sophistication involved in the each round of simulations will improve, in order to meet the requirements set in conjunction with the science analysis goals of the key project science working groups, and with the data reduction/pipeline development testing goals of the data management project. As part of this effort we are now producing large N-body simulations (109 − 1010 particles) in the MareNostrum machine at the Barcelona Supercomputer Center. 8 Conclusions The Dark Energy Survey will employ four complementary techniques to study dark energy: galaxy clusters, weak lensing, galaxy clustering and supernova distances. The statistical reach of these techniques is well understood; in the DES, each of them will deliver statistical constraints on dark energy (see 200 F. J. Castander and the Dark Energy Survey Collaboration Table 1) that are stronger than the best combined constraints available today [19, 20, 18]. The four techniques are both geometrical and growth rate techniques offering the possibility of separating different possibilities of the nature of dark energy. Moreover, our collaboration is making substantial progress towards identifying and understanding the dominant astrophysical uncertainties and observational systematic errors for each of these methods and one of our important goals is to further explore and develop methods to control these systematic errors. DES as a large scale mid-term survey is a logical step towards the more ambitious projects of the future. DES will use DECam, a powerful new wide-field survey instrument at the Blanco 4m. As a relatively shallow survey, the DES makes use of source galaxies that are large enough to be resolved and bright enough so that their photometric redshifts can be well calibrated. The collaboration institutions have a proven record in astronomical data management and have the capacity to manage large data sets. DES will also provide the astronomical community with a legacy survey of the largest volume of the universe sampled up to now. Acknowledgements: FJC acknowledges support from the Spanish Ministerio de Educación y Ciencia, project AYA2005-09413-C02-01 with EC-FEDER funding and from the Generalitat de Catalunya, project 2005SGR00728. References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. T. Abbott et al: astro-ph/0610346 J. Annis et al: astro-ph/0610194 J. Annis et al: astro-ph/0610195 A. Cooray et al: ApJ 557, L7 (2001) D. Eisenstein et al: ApJ 633, 560 (2005) P. Foslaba, E. Gaztañaga and F. Castander: ApJ 597, L89 (2003) M. Hamuy et al: AJ 112, 2438 (1996) S.E. Holland et al:SPIE 50(3), 225 (2003) W. Hu: Phys.Rev.D 67, 081304 (2003) W. Hu: Phys.Rev.D 66, 083515 (2002) D. Huterer: Phys.Rev.D 65, 063001 (2002) S. Majumdar & J.J. Mohr: ApJ 613, 41 (2004) S. Perlmutter et al: ApJ 483, 565 (1997) S. Perlmutter et al: ApJ 517, 565 (1999) A.G. Riess, W.H. Press and R.P. Kirshner: ApJ 473, 88 (1996) A.G. Riess et al: AJ, 116, 1009 (1998) J. Ruhl et al: SPIE 5498, 11 (2004) U. Seljak et al: Phys.Rev.D 71, 103515 (2005) D.N. Spergel et al: ApJS 148, 175 (2003) M. Tegmark et al: Phys.Rev.D 69, 103501 (2004) Observational cosmology at high redshift A. Fernández-Soto Departamento de Astronomı́a y Astrofı́sica, Universitat de València, alberto.fernandez@uv.es Summary. I offer a brief review of the evolution and present status of our observational knowledge of the high-redshift Universe. In particular, I focus on the different methods that have been devised to select distant objects, and the observational evidence in hand to support (or else) the standard evolutionary scenario. 1 How high is high? The study of objects at cosmological distances from us started in the sixties with the discovery and identification of quasi-stellar radiosources. The explanation of the features observed in the optical spectra of these objects as highly redshifted hydrogen lines opened the door to the very distant Universe. For the next thirty years after the discovery of quasars, the early Universe was the realm of monsters that could only be observed at such large distances because of their peculiar properties. These included powerful radioemission (usually associated to jets), or the presence of a strong X-ray flux. Of course, these selection effects led to strong biases in all the census of high-redshift objects that were produced, a lack that was well known by astronomers. It was suggested by different groups in the eighties that different techniques, based on colour selection, could in fact overcome those selection effects [13, 16]. The idea had been around since [2], [21] and [20], and took momentum specially with the work by Steidel and collaborators, who designed a colourselection technique able to sieve objects at redshift z ≈ 3 almost routinely, via deep imaging through three filters. The next impulse came with the Hubble Space Telescope observations of the Hubble Deep Field in 1995. These observations, designed to be extremely deep and rich in colour information, opened a new era of discovery for high-redshift, normal galaxies at all redshifts out to z ≈ 6 (see, for example, [14, 7]). Next step, the same selection techniques were applied once more to “monsters” like very luminous quasars, which, being extremely scarce in terms of number density in the sky, can only be detected with very large surveys that 202 A. Fernández-Soto Fig. 1. Record-breaking most distant galaxies and quasars as a function of time. Observe the different alternatives that happened over the nineties, as well as the breaks in both curves due to new techniques or the arrival of deeper surveys. include a wealth of colour information. This has been the case for the Sloan Digital Sky Survey, and this way the most distant quasars have been detected [6], out to z ≈ 6.5. At the time of writing this article, as can be seen in Figure 1, quasars and galaxies fight for the “universal record” of the most distant object detected, while a third team has joined the competition. Recent detections and fast follow-up of gamma ray burst afterglows have allowed for the analysis of transient sources out to redshift z ≈ 6.3. With all the data in hand, there is a picture that is beginning to emerge. Early star formation happened much faster in the history of the Universe than was thought only 15 years ago. The same celerity applies to the growth of the first galaxies, which were in place when the Universe was only 10% of its present age. Some of these observations do fit in our general view of cosmological evolution, while others seem to pose some problems, at least for the most nave of our models. 2 Hydrogen absorption in the IGM Thanks to the analysis of the absorption lines detected in the spectra of distant quasars we know that the space between the galaxies is not completely empty, but filled with gas (mostly hydrogen) that sometimes creates clumps that can be detected in absorption [23]. Hydrogen Lyman-alpha absorption is most easily detected, and because of the different redshift of the emitter and the absorbers, is imprinted on the spectrum of the background source as absorption lines at wavelengths bluer than the emitter observed-frame Lyman- Observational cosmology at high redshift 203 Fig. 2. Spectra of three quasars at redshifts 1.3, 3.6 and 6.3. The effect of the intergalactic medium absorption bluewards of the Lyman-alpha emission line can be seen, combined with the progressive reddening of the spectra due to the cosmological redshift. alpha. Moreover, the imprint of the densest clouds is also seen on the spectra as a complete absorption below the Lyman-limit in the absorber frame. It is a well-known fact that the number density of these intervening absorbers grows with redshift, in such a way that the flux of the highest-redshift quasars can be completely obscured below the Lyman-alpha line due to the sum of all the individual absorbers. Figure 2 shows the spectra of three different quasars at redshifts z ≈ 1, 3.5, and 6, where this effect is clearly seen. The net effect of the intergalactic medium is, hence, the progressive obscuration of the detected radiation in the bluer bands. It proceeds in such a way that objects at redshift z ≈ 2.5 are strongly absorbed in the U passband, while completely unabsorbed in the BV RI bands. Similarly, objects at z ≈ 3.5 are completely invisible in U , strongly absorbed in V , and unabsorbed at redder wavelengths. These are features that can easily be singled out in colour-colour diagrams, and served Steidel and collaborators to start their program to detect high-redshift galaxies and revolutionised the field (see, for example, [25]). Figure 3 shows two examples of this kind of analysis. However, this kind of colour-colour selection techniques have recently been challenged by the results of the VLT-Virmos Deep Survey [15]. This survey, the first one ever which has been able to produce a large, deep, and complete spectroscopic survey, has found that whereas the colour-selected samples are indeed very pure (very few selected objects are not in the expected redshift range),they are very far from being complete (see Figure 4). In fact, the VVDS group finds that up to a factor of five more galaxies are found in their com- 204 A. Fernández-Soto Fig. 3. LEFT: Upper panel: Unabsorbed and absorbed spectra of a galaxy at redshift z = 3, plotted together with the response functions of the HST filters through which the Hubble Deep Field was observed. Lower panel: U BV I images of a z = 2.8 galaxy in the Hubble Deep Field (adapted by the author from [8]). RIGHT: Upper panels: Images taken through the BRI filters of a fragment of the NOAO Deep Wide Field Survey. Lower panel: Filter responses plotted together with model spectra of z = 3 and z = 4 galaxies (from [10]). plete survey than would be selected using the colour-colour technique. This result would imply that most of the results about the high-redshift Universe that have been presented over the last few years could actually be biased. Luminosity functions, star formation histories, clustering measurements, could in fact correspond only to the tip of the iceberg. 3 Other selection techniques Some groups have tried to detect and select high-redshift objects using other properties different than colour. For example, Roser Pelló and her group have used the magnification power of large galaxy clusters to peer behind them, the expectation being that those galaxies that lie on positions close to the cluster caustic will be magnified by a factor of 20 or more, which will render visible some normal galaxies at distances far beyond what our telescopes would usually detect. Some spectacular results can be seen in [22]. Yet other groups perform Lyman-alpha emission searches, obtaining very deep images of random fields through a narrow-band filter and a broad-band filter that encompasses the former. An object that shines bright in the narrowband image but is less bright or even absent in the wide-band must have an emission line at that precise wavelength. For example, the Japanese groups working at Subaru have produced in this way samples of galaxies at z = 4.86 Observational cosmology at high redshift 205 Fig. 4. Colour-colour diagram as measured by the VVDS group. The region of the diagram that would be selected by the usual techniques is marked with a dashed line. It can be seen that many of the spectroscopically-detected high-redshift objects lie outside the region (from [15]). –1 ergs s cm A (1+z) ) ([19], Lyman-alpha at 7110 Å), z = 5.71 ([1], Lyman-alpha at 8160 Å), and z = 6.55 ([12], Lyman-alpha at 9210 Å). Figure 5 (taken from [12]) shows the images and spectra of two of the z = 6.55 galaxies. J132415.7+273058 (a) J132418.3+271455 (b) combined (c) 0 10 5 Flux Density (10 –19 –1 –2 –1 10 0 10 3–pix smoothed 5 0 1210 1220 Wavelength (A) Fig. 5. Narrow- and wide-band images and spectra of two z = 6.55 galaxies detected through the Subaru emission-line search (from [12]). 206 A. Fernández-Soto The colour technique, in a slightly different flavour, has also been applied to the selection of very high-redshift gamma ray burst afterglows. Nowadays it is possible to quickly react to trigger signals sent by orbiting satellites (Swift being the state-of-the-art in that field) and use large ground-based telescopes to study them before they fade away in only a few days. In this manner [26] detected and (photometrically, see Figure 6) singled out GRB050904 at z = 6.28, later observed with Subaru by [11]. Fig. 6. Spectral energy distribution of GRB050904, and best fit z = 6.3 spectrum as determined by the author. The photometric measurements include those taken by the UVOT telescope on board Swift [3], and those presented in [26], corrected to (t − t0 ) = 1.155 days. 4 The most distant objects: Open issues Over the last few years our image of the distant Universe has changed from a basically empty one to that of a place full of action. However, some problems remain open that can strongly affect our perception. In this section we will talk about some of those. 4.1 Extremely red objects As was mentioned above, the deepest surveys nowadays routinely select objects with peculiar colours, lending them particular attention. One family of objects that has been subject to considerable discusion is that of the extremely red objects (EROs). EROs are detected based on very red (I − K) or (R − K) colours, as high as 6 or more. [27] or [18] discuss the nature of some of them, Observational cosmology at high redshift 207 and find that the two standard scenarios can actually fit at least some of the observations: we are either detecting very luminous galaxies at extremely high redshifts (z ≈ 7 and beyond) or extremely reddened galaxies with both young and old stellar populations at intermediate redshifts (z ≈ 3). In either case, the standard model of galaxy evolution and growth must be stretched to include a large population of massive stars whose birth happened at extremely high redshift, in fact uncomfortably high in some cases. 4.2 Metal content According to all plausible models, the metal content of the Universe has been built up as time proceeded, from an almost pristine H+He mixture at the time of recombination to its present, very inhomogeneous, status. However, the best probe of the metal content that we can observe–the emission spectra of quasars–offers an incredibly boring scenario: the average quasar metallicity does not change between redshift 6 and redshift 2 [17]. It is certainly true that the quasar emission regions are far from typical, and cannot be taken as representative of any average universal value. The fact remains, though, as an uncomfortable thought. When the metallicity of quasar absorption lines is used instead, the buildup does indeed exist, although the scatter in the data– provoked by the large inhomogeneities expected in the density distribution of the Universe–is in fact almost as large as the measured change in metallicity. 4.3 Reionisation? Since the SDSS group discovered the first z > 5 quasars [4], it was noted that the absorption due to hydrogen close to the quasar redshift was very close to complete. The presence of a completely opaque trough at Lyman-alpha wavelengths, predicted by [9], would point at the presence of a large fraction of neutral hydrogen. This so-called Gunn-Peterson effect has not been detected out to z = 6, and this means that the Universe remains almost completely ionised out to that distance. The detection of the z ≈ 6.3 trough, as claimed by several authors (see for example [5]), would mean that the Universe did indeed reionise at some epoch close to that, and certainly earlier than z = 6. A more recent and detailed study by Songaila shows that the Hi opacity out to z = 6.5 can be smoothly linked to that at lower redshift, making that claim dubious (see Figure 6). This study has been challenged again by [6], who add new data that supports their previous result. It is certainly true, in any case, that our observing abilities must be approaching the epoch of reionisation, and that the determination of that moment, however fuzzy it must be, is one of the key ingredients in the history of galaxy formation. References 1. Ajiki, M., et al, AJ, 126, 2091 (2003) 208 A. Fernández-Soto Fig. 7. Opacity of the intergalactic medium at the Lyman-alpha wavelength as a function of redshift. The plot includes measurements of quasar spectra from z = 2.3 to z = 6.3 (taken from [24]). 2. Baum, W. A., IAU Symp. 15: Problems of Extra-Galactic Research, 15, 390 (1962) 3. Cucchiara, A., et al., GRB Coordinates Network, 3923, 1 (2005) 4. Fan, X., et al., AJ, 118, 1 (1999) 5. Fan, X., et al., AJ, 123, 1247 (2002) 6. Fan, X., et al., AJ, 132, 117 (2006) 7. Fernández-Soto, A., Lanzetta, K. M., & Yahil, A., ApJ, 513, 34 (1999) 8. Giavalisco, M., ARA&A, 40, 579 (2002) 9. Gunn, J. E., & Peterson, B. A., ApJ, 142, 1633 (1965) 10. Jannuzi, B. T., & Dey, A., ASP Conf. Ser. 193: The Hy-Redshift Universe: Galaxy Formation and Evolution at High Redshift, 193, 258 (1999) 11. Kawai, N., et al., Nature, 440, 184 (2006) 12. Kodaira, K., et al., PASJ, 55, L17 (2003) 13. Koo, D. C., AJ, 90, 418 (1985) 14. Lanzetta, K. M., Yahil, A., & Fernández-Soto, A., Nature, 381, 759 (1996) 15. Le Fèvre, O., et al., Nature, 437, 519 (2005) 16. Loh, E. D., & Spillar, E. J., ApJ, 303, 154 (1986) 17. Maiolino, R., et al., ApJ, 596, L155 (2003) 18. Mobasher, B., et al., ApJ, 600, L167 (2004) 19. Ouchi, M., et al., ApJ, 582, 60 (2003) 20. Partridge, R. B., ApJ, 192, 241 (1974) 21. Peebles, P. J. E., & Partridge, R. B., ApJ, 148, 713 (1967) 22. Pelló, R., et al., A&A, 416, L35 (2004) 23. Sargent, W. L. W., Young, P. J., Boksenberg, A., & Tytler, D., ApJSS, 42, 41 (1980) 24. Songaila, A., AJ, 127, 2598 (2004) 25. Steidel, C. C., et al., ApJ, 462, L17 (1996) 26. Tagliaferri, G., et al., A&A, 443, L1 (2005) 27. Yan, H., et al., ApJ, 616, 63 (2004) An Hα approach to the evolution of the galaxy population of the universe J. Gallego1 , V. Villar1 , S. Pascual1 , J. Zamorano1, K. Noeske2 , D.C. Koo2 , P.G. Pérez-González1, and G. Barro1 1 2 Departamento de Astrofı́sica, Universidad Complutense de Madrid, Avda. Complutense s/n 28040 Madrid, Spain, jgm@astrax.fis.ucm.es Lick Observatory, California, University of California, CA95060 Santa Cruz, USA Summary. We present a long-term project to study the characteristics and evolution of current star-forming galaxies at different redshifts. The selection of the samples is carried out in an homogeneous way by detecting Hα emission-line objects in deep images taken with narrow-band filters tuned to the corresponding wavelength for a given redshift. In this way we are providing a complementary approach to the large program surveys using ultraviolet or infrared data. Already targeted redshifts are z=0.24, z=0.40 and z=0.84. Here we present results from the three redshift bins, focusing into the cuasi-local z=0.24 regime and the z=0.84 regime, the last corresponding to an epoch of the Universe where most of the Star Formation Rate (SFR) activity is supposed to happen. First we analyze photometric redshifts, emission-line luminosities and multi-band photometry for a sample of Hα emission-line galaxies at z=0.24, selected in the 8200Å atmospheric window. The physical properties of these galaxies appear to be remarkably similar to those of local galaxies selected in the same way. Our results support the idea that the higher Star Formation Density measured at redshift z=0.24 is due to an increase of the density of bursting galaxies and not to an intrinsic change on overall galaxy properties. We present results from the analysis of a sample of Hα emission-line galaxies at z=0.84 selected by their contrast in a narrow band filter centered in the near-infrared 1.21µm region. The corresponding Hα luminosity function and infered SFR density, assuming an average +0.03 extinction of AHα = 1 and a concordance cosmology, is 0.12−0.02 M yr −1 M pc−3 . This value confirms an increase of ∼ 10 with respect to the local value. Combining just the Hα-based SFR densities obtained by our group from z∼ 0, z=0.24, z=0.4 and z=0.8, we obtain an evolution ∝ (1+z)β where β = 3.2 ± 0.7. Using 24µm fluxes as measured by Spitzer, we have estimated the Far-IR based SFRs. The ratio between the SFRs as traced by Far-IR and Hα increases with Far-IR luminosity above a given Far-IR luminosity threshold. Finally, we have used IRAC data to estimate the rest-frame K band luminosities and to infere stellar masses. The average stellar mass for Hα-selected star-forming galaxies at z=0.8 is M=5x1010 M . The average specific star formation rate for z=0.8 galaxies is about ×10 the local value for similar objects. 210 Gallego et al. 1 The extended-UCM Survey The Star Formation Rate (SFR) density of the Universe is one of the key observables needed for our understanding of galaxy evolution. Deep redshift surveys suggest that star-formation activity substantially increases with redshift until z ' 1 (see [8] for a nice review). Detailed theoretical works (e.g., [17, 18, 6, 1, 30, 13, 4]) are able to predict the global star formation history of the Universe, i.e. the comoving number density of galaxies as a function of SFR, and as a function of redshift. Studying the evolution of the SFR and the properties of the star-forming galaxy populations can thus provide important clues on galaxy formation and evolution. To test directly how substantial evolution in the star-formation activity has occurred we need to measure the Star Formation Rate (SFR) density of the Universe and the properties of the corresponding star-forming galaxy populations at different redshifts using similar techniques. Several tracers can be used to obtain SFRs: ultra-violet (UV) continuum, nebular lines such [Oii]λ3727 and Hα, and far infra-red (FIR) luminosities. However significant discrepancies have been found when comparing the values obtained from these tracers, due to dust extinction, metallicity and different spatial origins of the emission (See Figure 1). The studies of local galaxies reveal a discrepancy between SFRs obtained with UV and Hα luminosities compared with those obtained from FIR luminosities. This discrepancy seems to increase with the SFR. At higher redshifts the inconsistency between tracers is also present. [28] studied a sample of 12 galaxies with redshifts from 0.4 to 1.4, detected with ISOCAM in the Hubble Deep Field South (HDF-S). They found the same discrepancy but they assumed an average dust extinciton correction factor of 4 for SFR(Hα). [5] studied a sample of 7 galaxies, 4 with z'0.4 and 3 with z'0.8 detected with ISOCAM in the Hubble Deep Field (HDF). In this case the Balmer decrement could be measured and the effect of dust extinction was fine corrected. They find that for a sub-sample at z'0.4 the SFR tracers were consistent. For a sub-sample at z'0.8 they find that the SFRs derived from UV, [Oii]λ3727 and FIR luminosities are, respectively, lower, similar and higher, than the results obtained from Hα. The authors also find a relation between SFRIR /SFRHα ratio and IR luminosity. A possible explanation is that opaque dust clouds are present in the star forming regions. [9] have studied a sample of 16 distant galaxies detected by ISOCAM. They find that galaxies with SFRs over 90 - 130 M /yr present a SFRIR /SFRHα ratio up to 2.5. Our goal is to build homogeneous samples of current star-forming galaxies covering the redshift range from z∼1 to the current epoch. The objects are selected by their flux excess in a narrow-band filter centered at the wavelength corresponding to Hαλ6563Å redshifted at the target redshift. At this moment, optical filters at 8200Å and 9200Å have been used to study the z=0.24 and z=0.4 universes. Work at 11810Å is in progress to study the z=0.84 universe. The Hα-based SFR of the Universe 211 Fig. 1. Hα and Star Formation Rate density function of the Universe at z=0.24 as estimated by [25]. Different symbols correspond to different observing runs at Calar Alto 2.2m telescope (CAHA00 and CAHA01) and La Palma 2.5m Isaac Newton Telescope (INT01). Thick line is the Schechter fit to z=0.02 ([10]), thin line is the Schechter fit to z=0.24 by [31] and dotted line is the fit to z=0.40 from [25]. All ofthem are based on Hα fluxes corrected by an average extinction of A(Hα)=1. 2 Overall properties of Hα emitters at z=0.24 Here we analyze the sample of candidates to emission-line galaxies used by [24] to estimate the Hα-based SFR density of the Universe at z=0.24. In that paper, we presented deep images taken with the 2.2m telescope of Calar Alto Observatory (Almerı́a, Spain) on the ELAIS-N1 region using a narrow band filter centered in 8200Å with a FWHM of 160Å and a broad-band filter to probe the continuum. This specific narrow-band filter was selected because at 8200Å there is a substantial gap in the sky-night OH emission lines. The broad-band was selected because it has almost the same effective wavelength. Deep multi-band photometry for our sample was obtained from the INT Wide Field Survey (hereafter INT-WFS). Our field was imaged with five filters (Ugriz ), ranging from 3664Å (U filter) to 8953Å (z filter). Limiting magnitudes were ∼22.9 in the broad band and ∼21.0 in the narrow band. There were 61 objects in the original sample. 16 of them were marginal detections, 212 Gallego et al. −1 log SFR(Hα) [MJ yr ] 0 1 2 0 log Φ [Mpc−3] −1 −2 −3 −4 −5 41 42 43 −1 log L(Hα) [erg s ] Fig. 2. Hα luminosity and SFR density function for our sample of Hα-selected starforming galaxies at z=0.84. The extinction correction applied is A(Hα=1 mag. The contamination by [NII]λλ6548,6584 has been subtracted using average values. Filled circles correspond to the WHT2003 data. Open circles correspond to the CAHA3.5m 2004. Thin line is the Schechter fit to the z=0 sample from [10]. The dash thin line is the Schechter fit to z=0.2 from [24]. Black line is the Schechter function for [32] at z=0.7. Thick dashed line is the corresponding to [14] sample. Finally the red line is the Schechter fit to our z=0.8 data. See Table 1 for more details. detected only in the narrow band filter. Stars were flagged using the stellarity parameter CLASS STAR given by Sextractor ([2]). In order to discriminate between genuine Hα emitters and potential sources of contamination, i.e. stars and other emission line objects ([OIII]5007 at z∼0.64, Hβ at z∼ 0.68 or [OII]3727 at z∼ 1.2), photometric redshifts were computed through a standard minimization procedure, using the public code hyperz ([3]). Based in both the stellarity and photometric redshift, 13 objects were finally classified as stars and 33 objects were confirmed as Hα emitters at z=0.24. We have found 6 candidates to be galaxies at z∼0.6 and 3 at ∼1.2. The presence of these objects at these relatively bright fluxes implies a stronger evolution of the population of [OIII]5007 and [OII]3727 emitters than the corresponding to Hα. We find a total extinction-corrected Hα luminosity density of (4.6±0.3) 1039 erg s−1 Mpc−3 at z=0.24. Assuming a constant relation between the Hα The Hα-based SFR of the Universe Reference (1) Yan et al. (1999) Hopkins et al. (2000) Tresse et al. (2002) This work ∆z (2) 0.7<z<1.9 0.7<z<1.8 0.5<z<1.1 0.79<z<0.85 z̄ (3) 1.34 1.3 0.73 0.84 N (4) 33 37 33 122 log L∗ (5) 43.24 43.30 42.37 42.16 log φ∗ (6) -2.77 -3.07 –2.39 -1.91 α (7) -1.35 -1.6 -1.31 -1.3 213 Selection (8) SLS, Hα SLS, Hα I band NB, Hα Table 1. Schechter parameters for the available Hα-based luminosity functions at z ∼ 1 (see Figures 2 and 3). All of them are translated to a Einstein-de Sitter cosmology with H0 =50 km−1 Mpc−1 . Col.(1): Published reference. Col.(2): Redshift range covered. Col.(3): Average redshift. Col.(4): Total number of galaxies. Col.(5), (6), (7): Schechter parameters. Col.(8): Selection criteria used, where SLS corresponds to slit-less spectroscopy and NB corresponds to narrow-band filters. luminosity and star formation rate, the SFR density in the volume covered is (0.036±0.002) M yr−1 Mpc−3 . This value is a factor of ∼ 3.8 higher than the local SFR density and consistent with the strong increase in the SFR density from z=0 to z=1 (see Figure 1 and Figure 2). Hα emitters at z=0.24 exhibit properties similar to the local analogs, in distribution of Hα fluxes and equivalent widths. The downsizing scenario ([7]), in which rapid star-forming galaxies evolve smoothly in luminosity with decreasing redshift, seems to be a possible model to explain the evolution of these galaxies. 3 The Hα-based SFR density at z=0.84 We present results from the analysis of a pilot sample of Hα emission-line galaxies at z=0.8 selected by their contrast in a narrow band filter centered in the near-infrared 1.21µm region. The data was collected using two observing configurations: A wide-field search was performed using the near-infrared camera OMEGA-20001 at the 3.5m telescope in the Calar Alto Observatory (Almerı́a, Spain). OMEGA-2000 is equipped with a 2k×2k Hawaii-2 detector with 18µm pixels (0.45arcsec on the sky, total area 15’×15’ per pointing). Two pointings were obtained in the Extended Groth Strip in april 2005 and may 2006. The total exposure times were ∼7200s and ∼18000s in the J and Jc filters (centered at 1.20µm and corresponding to Hα at z=0.84) per pointing. The average seeing was ∼ 1.1arcsec. The average limiting flux was 8×10−17 erg cm−2 s−1 . A second deeper search was done in an smaller field of view (13.7 sqr. arcmin) using the near-infrared camera INGRID ([23] ) at the 4.2m William Herschel Telescope in the Roque de los Muchachos Observatory (La Palma, Spain). One single pointing was performed in the Groth region in May 2003. 1 http://www.mpia-hd.mpg.de/IRCAM/O2000/index.html Gallego et al. 0.1 0.2 1 2 -0.5 -3 [erg s Mpc ] 40.6 -1 40.2 -1 log SFRd(Hα) [Msun y-1 Mpc-3] 0 Redshift 0.5 39.8 log Ld(Hα) 214 -1.5 39.4 -2 39 0 2 4 6 8 10 Lookback time (Gy) Fig. 3. Evolution of the SFR density of the Universe as measured using Balmer recombination lines. Different colors depend on the selection criteria used. In blue are those samples where Hα fluxes were used for selection. In black the selection was done using a broad band apparent flux limit. In red we have plotted those densities obtained by the extended-UCM survey selecting galaxies by their Hα fluxes. Red filled triangle is from [10]. Blue cross is for [12]. Black empty squares are [16]. Blue closed rhomb is [31]. [24] and [25] are red filled pentagons. Blue filled rhomb is [32]. This work is the red pentagon at z=0.8. Black filled square is 1999MNRAS.306..843G. [14] is the pink filled triangle. Black open circle is Yan99. Pink filled circle is [15]. The blue open triangle is [21]. Finally, the green triangle is [27] measured using Hβ fluxes. The total exposure times were 11460s and 13400s in the J band and in an ucm-designed narrow-band filter centered at 1.18µm (corresponding to Hα at z=0.80). The average seeing was ∼ 1.3arcsec and ∼ 1.1arcsec. This pointing is covered by one ofthe O-2000 pointings. Limiting flux was 1.4×10−17 erg cm−2 s−1 . The total area covered by all the pointings is 415 square arcmin. The corresponding Hα luminosity function (see Table 1) and infered SFR density is 0.12Myr−1 M pc−3 . This value confirms an increase of ∼ 10 with respect to the local value (see Figure 2). Combining just the Hα-based SFR densities obtained by our group from z∼ 0, z=0.24, z=0.4 and z=0.8, we obtain an evolution ∝ (1+z)β where β = 3.6 (see Figure 3). The Hα-based SFR of the Universe 215 The final sample of emission-line galaxies used for this paper was 122 objects. Among the fraction of objects with HST imaging available, most of them show a disky morphology. Spectroscopic confirmation came out from the redshifts obtained by the DEEP2 Galaxy Redshift Survey ([19]) using Keck telescopes. 55 candidates (45% of the total) have an accurate spectroscopic redshift, most of them confirming our interpretation. As a summary, 80% of the candidates are genuine emission-line galaxies at the right redshift, whereas 20% turn out to be galaxies at other redshifts. The Hα luminosity function was computed using a traditional V/Vmax method where the volumes were computed following the method explained by [26]. These authors compute compute a different volume for each object. This volume is determined by the intersection between the selection curve and the curve defined by the position of the emitter as a function of the redshift allowed by the filter transmitance in the color-magnitude selection diagram The resulting Schechter best fitting parameters are: α − 1.33 ± 0.20, φ∗ = 10−1.89±0.09 Mpc−3 , and L∗ = 1042.08±0.06 erg s−1 . 40 The corresponding total Hα luminosity density is (2.1+0.5 erg s−1 −0.4 ) ×10 +0.04 −3 −1 −3 Mpc . The infered SFR density is 0.16−0.03 M yr M pc . No correction for AGN contamination was considered, given the uncertainty classifying AGNs using the data available for our sample. Deep X-ray data for the Groth field was obtained by Chandra ([22]), but only two of our sources were detected. An idea of the amount of correction by AGNs can be provided by the local UCM survey. For this sample, AGNs contribute with 8% in number density and 15% in Hα luminosity density. The Hα-based SFR density obtained confirms an increase of ∼ 10 with respect to the local value. Combining just the Hα-based SFR densities obtained by our group from z = 0.02, z = 0.24, z = 0.40 and z = 0.84, we obtain an evolution ∝(1+z)β where β = 3.2 ± 0.7. 4 Spitzer data for star-forming galaxies at z=0.84 Using 24µm fluxes as measured by Spitzer, we have estimated the Far-IR based SFRs. We find that Hα-selected galaxies also show the trend previously already found for Far-IR selected galaxies. The ratio between the SFRs as traced by Far-IR and Hα increases with Far-IR luminosity above a given Far-IR luminosity threshold. However, only when a larger sample becomes available we will be able to better address this issue. Finally, we have used a combination of IRAC 3.6µm and 4.5µm to estimate the rest-frame K band luminosities. Assuming a constant M/L ratio, they allowed us to infere stellar masses. The average stellar mass for Hα-selected star-forming galaxies at z=0.8 is M=5x1010 M . This is about ×5 larger than the tipical stellar mass estimated for our local reference sample. In consequence, the average specific star formation rate is about ×10 the local value for similar objects. 216 Gallego et al. References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. Baugh, C. M.; Cole, S.; Frenk, C. S.; Lacey, C. G., 1998, ApJ 498, 504 Bertin, E. & Arnouts, S. 1996, A&ASS 117, 393 Bolzonella, M., Miralles, J., & Pelló, R. 2000, A&A 363, 476 Calura, F., Matteucci, F., 2003, ApJ 596, 734 Cardiel, N., Elbaz, D., Schiavon, R.P., et al., 2003, ApJ 584, 76 Cole, S.; Aragon-Salamanca, A.; Frenk, C., et al., 1994, MNRAS 271, 781 Cowie, L., Songaila, A., Hu, E., & Cohen, J. 1996, AJ 112, 839 Ferguson, H.C., 2000, ARA&A 38, 667 Flores, H., Hammer, F.,Elbaz, D., et al., 2004, A&A 415, 885 Gallego, J., Zamorano, J., Aragón-Salamanca, A., et al., 1995, ApJ 455, L1 Glazebrook, K.; Blake, C.; Economou, F.et al., 1999, MNRAS 306, 843 Gronwall, C., 1999, AIPC 470, 335 Hoeft, M.; Mucket, J.P.; Heide, P., 2002, ApJ 570, 33 Hopkins, A.M., Connolly, A.J., Szalay, A.S., 2000, AJ 120, 2843 Iwamuro, F.; Motohara, K.; Maihara, T., et al. 2000, PASJ 52, 73 Jones, D.H.; Bland-Hawthorn, J., 2001, ApJ 550, 593 Kauffmann, G.; White, S., Guiderdoni, B., 1993, MNRAS 264, 201 Kauffmann, G.; Colberg, J.M.; Diaferio, A.; White, S., 1999, MNRAS 303, 188 Koo D.C., 2002, AAS 34, 1320 Madau, P., Pozzetti, L., & Dickinson, M. 1998, ApJ 498, 106 Moorwood, A. F. M.; van der Werf, P. P.; Cuby, J. G., et al., 2000, A&A 362, 9 Nandra K., Laird E.S., Adelberger K., et al., 2005, MNRAS 356, 568 Packham C., Thompson K.L., Zurita A. et al., 2003, MNRAS 345, 395 Pascual, S., Gallego, J., Aragón-Salamanca, A., & Zamorano, J. 2001, A&A, 379, 798 Pascual, S., 2004, Ph.D. Thesis, Universidad Complutense de Madrid Pascual S., Gallego J., Zamorano J., 2006, PASP in press, astro-ph/0611121 Pettini, M.; Kellogg, M.; Steidel, C., et al., 1998, ApJ 508, 539 Rigopoulou, D., et al., 2000, ApJ 537, L85 Rowan-Robinson, M., & Elais Consortium. 1999, ESA SP-427: The Universe as Seen by ISO, 427, 1011 Somerville, R.; Primack, J.; Faber, S.M., 2001, MNRAS 320, 504 Tresse, L. & Maddox, S.J., 1998, ApJ 495, 691 Tresse, L.; Maddox, S., Le Fèvre, O.; Cuby, J., 2002, MNRAS 337, 369 Yan, L.; McCarthy, P.J., Freudling, W., et al., 1999, ApJL, 519, 47 Session V The Galaxy and its components Multi-wavelength Astronomy and the unidentified γ-ray sources J. Martı́-Ribas Departamento de Fı́sica, Escuela Politécnica Superior, Universidad de Jaén, Campus Las Lagunillas s/n, A3-420, 23071 Jaén, Spain, jmarti@ujaen.es Summary. The problem of unidentified sources has been a recurrent one in the history of Astronomy. Soon after the opening of a new spectral window, the first objects detected through it are often poorly located and very difficult to associate with counterparts seen at other more familiar wavelengths, such as the optical band. As an example of this statement, we can recall the early times of Radioastronomy nearly half a century ago. In a historical paper by Baade & Minkowski[5], we can read: although the sources in Cassiopeia and Cygnus A are among the brightest and earliest-known radio sources of the sky, all attempts to identify them with astronomical objects in the visible range have failed so far. Some decades later, the unidentified source problem vanished thanks to the technical development of radio interferometers. Ironically, these instruments working at radio wavelengths provide today the most accurate positions of celestial bodies. The problem shifted to the domain of X-rays in the 60s and 70s of last century. Again, the technical progress solved it once more. Today X-ray observatories, such as Chandra or XMM, are able to deliver positions one to few arc-second accurate, thus becoming comparable to those from ground based optical telescopes. At present, a significant number of unidentified sources exist in the relatively young branch of γ-ray astronomy where the state-of-the art instruments are not yet able to deliver point spread functions (PSF) below 0.1 degrees. This review will discuss the limitations and strategies that we are currently facing to find out the nature of celestial γ-ray sources. 1 Past, present and future γ-ray observatories γ-ray telescopes trace some of the most energetic processes in the Universe and several physical mechanisms are know to produce them both in continuum and line emission. Among the most relevant ones in the astrophysical context, we can distinguish the so called leptonic and hadronic processes. Leptonic γ-rays originate mainly by the inverse Compton effect of different populations of seed photons by relativistic electrons. Instead, γ-rays of hadronic origin result of neutral pions produced after the collision of relativistic protons with hydrogen 220 J. Martı́-Ribas nuclei of the interstellar medium (ISM). The production of energetic particles required to trigger γ-ray emission is often connected with the existence of compact and relativistic objects, such as neutron stars and black holes. Their compactness is the reason why celestial γ-ray emission can be so violent and variable, as compared to other less energetic domains of the electromagnetic spectrum. The reader is referred to dedicated monographs such as [23] for a detailed view on the fundamentals of γ-ray astronomy. 1.1 Observatories in space The shielding properties of the Earth atmosphere force most γ-ray detectors to work on board satellite observatories in space. During the last decade, the COMPTON Gamma Ray Observatory (GRO) was one of these satellites and its legacy became seminal for our understanding of the γ-ray sky. The COMPTON-GRO carried four instruments covering the 30 keV to 30 GeV range. Among them, the Energetic Gamma-ray Experiment Telescope (EGRET) produced a remarkable catalogue of 271 high energy (E ≥ 100 MeV) γ ray sources[15]. A significant fraction of them are identified with blazars and are, therefore, extragalactic. The other identified sources are of galactic origin and include several pulsars, a solar flare, the Large Magellanic Cloud, a probable radio galaxy (Cen A) and two γ-ray binaries. However, as of today, nearly half of the EGRET catalogue still remains unidentified and this is mostly due to its poor location accuracy (∼ 1◦ ). This difficulty is illustrated in Fig. 1 where a representative example of an EGRET source is shown. Ignoring the galactic or extragalactic nature of about 50% of all high energy γ-ray sources in the sky is indeed a challenging situation that needs to be solved. At present, the γ-ray sky is being monitored by the International GammaRay Astrophysics Laboratory (INTEGRAL), carrying different instruments sensitive in the 3 keV to 10 MeV range. One of the most important contributions of this satellite is the regular INTEGRAL Galactic Bulge Monitoring performed every few days with a complete hexagonal dither pattern with the ISGRI (20-150 keV) and JEM-X (3-25 kev) instruments. This is a valuable service to the community that has unveiled already many new soft γ-ray sources. Updated status, light curves and images are quickly available on-line1 . By late 2007, the γ-ray community is eagerly expecting the launch of the Gamma Ray Large Area Space Telescope (GLAST) satellite. This observatory will carry a wide field large area telescope sensitive within 100 MeV and 300 GeV. Despite of being a relatively small mission, this observatory is likely to produce a major progress in γ-ray astronomy mainly due to its expected source source location capability below one arc-minute. Therefore, the problem of unidentified EGRET sources should be in a course towards solution in a few year horizon. 1 http://isdcul3.unige.ch/Science/BULGE/ Multi-wavelength Astronomy and the unidentified γ-ray sources 221 -10 30 DECLINATION (J2000) -11 00 30 -12 00 19 08 07 06 05 04 RIGHT ASCENSION (J2000) Cont peak flux = 4.8270E-01 JY/BEAM Levs = 5.000E-04 * (-4, 4, 6, 8, 16, 32, 64, 128) 03 02 Fig. 1. The EGRET source 3EG J1904-1124 and the 99, 95, 68 and 50 percent confidence contours of its location in the sky. Note that the field covers more than 1◦ across. The background objects are radio sources from the NRAO VLA Sky Survey (NVSS) at the 20 cm wavelength. Many of them are contained within the EGRET contours, thus rendering the identification a very difficult task. 1.2 Observatories in the ground At very high energies (VHE) of ∼ 1 TeV, γ-rays can be detected indirectly from the ground. Photons so energetic initiate electromagnetic cascade showers when colliding with atoms in the high Earth atmosphere. Electrons and positrons created in this way radiate Cherenkov light, within a cone angle of ∼ 1◦ , while moving with velocities higher than the local light velocity c/n, where n is the refractive index of the air just slightly above one. Cherenkov photons are able to penetrate the atmosphere and come in very short flashes, typically lasting a few ns. The their spectral peak is around the blue optical domain. The appearance of a γ-ray point source seen in Cherenkov light is spoiled into a diffuse object ∼ 1◦ across due to Coulomb scattering of particles in the cascade. Despite of this problem, different ground based telescopes and detectors equipped with fast photo-multipliers at their focus have been 222 J. Martı́-Ribas specially built to study VHE γ-ray sources in Cherenkov light. Their working principle is illustrated in Fig. 2. VHE GAMMA−RAY PARTICLE SHOWER 1 10 km CONE OF CHERENKOV LIGHT Fig. 2. Sketch of the working principle of an array of Cherenkov telescopes for the detection of very high energy γ-rays entering the Earth upper atmosphere. These so called Cherenkov telescopes must also face the huge problem of background rejection since cosmic rays also induce Cherenkov cascades unrelated to cosmic γ-ray sources. Indeed, it is estimated that a 10−4 fraction of the night sky background is of Cherenkov origin. Using coincidence techniques in arrays of Cherenkov telescopes is a very efficient way to get rid of most of the cosmic ray contribution. The first γ-rays of celestial origin detected in this way were reported by [11] and the Cherenkov technique has been continuously maturing since then. Today, the High Energy Stereoscopic System (H.E.S.S.) and the Major Atmospheric γ Imaging Cherenkov (MAGIC) are among the most sensitive Cherenkov facilities currently operating. They provide today the sharpest PSFs in modern γ-ray astronomy, of about 0.◦ 1, and therefore a source location capability of a few arc-minute depending on the statistics. The reader is referred to the respective web pages of these collaborations for Multi-wavelength Astronomy and the unidentified γ-ray sources 223 further details2 . The current trend in Cherenkov astronomy is towards building larger collecting dishes and/or larger number of telescopes in Cherenkov arrays. 2 Unidentified sources across the γ-ray band The census of unidentified sources sources is significantly different depending of the energy range being considered. At low-energy γ-rays where the INTEGRAL satellite operates, a total of ∼ 102 objects remain without a clear identification despite the efforts of many authors (see e.g. [19]). At high energy γ-rays, the domain of the old EGRET instrument, the total number of unidentified sources is of the same ∼ 102 order of magnitude. Here, the identification progress proceeds at a much slower pace due to the large positional error of EGRET. Finally, at the VHE domain of the MAGIC and H.E.S.S. telescopes, the number of unidentified sources is much smaller, of order ∼ 10, simply because there are not many TeV sources detected in the sky yet. In all cases, however, the total number of unidentified objects represents a significant fraction (∼ 50%) of the total emitters known in the respective energy bands. In order to facilitate the search, several γ-ray observing facilities and other international collaborations keep the basic information about unidentified sources available on line and updated. Among them, we can quoted here the list of INTEGRAL sources3, the H.E.S.S. source catalogue4 and the web page of the MINE collaboration5. Identifying a γ-ray source is, in most cases, like finding a needle in a haystack. Ideally, a reliable identification should require[13]: • A good positional coincidence of the γ-ray source with the proposed counterpart. • The counterpart candidate has to provide a viable γ-ray emission mechanism (e.g. being an AGN, microquasar, GMC, SNR, PWN, etc.). • A consistent fit within a multi-wavelength scenario (radio, IR, optical, X-rays, etc.). • A morphological match (in case of extended sources). Unfortunately, most identifications very often do not fully satisfy all these requirements. The best identified and most intensively observed γ-ray source is, without doubts, the Crab Nebula supernova remnant or M1. This is a very important source in γ-ray astronomy known since decades ago and being used as a primary calibrator by virtually all γ-ray telescopes. In contrast, the most mysterious γ-ray source currently known is probably TeV J2031+4130. This 2 3 4 5 http://magic.mppmu.mpg.de/introduction/index.html http://www.mpi-hd.mpg.de/hfm/HESS/HESS.html http://isdc.unige.ch/∼rodrigue/html/igrsources.html http://www.mpi-hd.mpg.de/hfm/HESS/public/HESS catalog.htm http://elbereth.obspm.fr/∼fuchs/mine.html 224 J. Martı́-Ribas is an extended object discovered some years ago with the stereoscopic High Energy γ Ray Astronomy (HEGRA) array of imaging Cherenkov telescopes in the direction of the Cygnus OB2 star association[2]. Despite years of intensive search, no counterpart candidate has been found a lower energies thus remaining as a completely unidentified source. Interestingly, TeV J2031+4130 could be the prototype of a new class of extended γ-ray sources since a similar case also exists in the southern hemisphere, namely HESS J1303−631[4]. Given the lack of detection of a low energy counterpart it has been suggested that hadronic processes instead of leptonic ones could be behind these extended TeV emitters. The proposed counterparts of common γ-ray sources can often be claimed with degrees of confidence intermediate between that of M1 (practically full certainty) and TeV J2031+4130 (full ignorance). 3 Identification strategies or how to find the needle in a haystack Very often, the identification of a γ-ray source with a significant positional uncertainty starts when a peculiar object is discovered at lower energies inside its 90% confidence error box. Here, peculiarity means often a variable object, an object with prominent emission lines, a conspicuous X-ray source inside the error box, etc., thus rendering it as a promising counterpart candidate. This would satisfy the first of the identification requirements quoted in previous section. Soon after this initial step, the peculiar object is heavily scrutinized within a multi-wavelength approach. These follow up observations are carried out, often in a coordinated way, in order to test how many additional requirements are also satisfied. If they do are fulfilled, then a counterpart identification can be claimed with confidence. This process can be very slow and painful, taking years to acquire all the necessary data and reach an acceptable theoretical understanding. A good example of such successful identifications are the so called γ-ray binaries LS 5039[21], LSI+61◦303[1] and PSR B1259−63[3]. These binary systems stand out inside their respective EGRET, MAGIC or H.E.S.S. error boxes due to a combination of the peculiarities quoted above. The remarkable position improvement provided by ground based Cherenkov telescopes has been a key factor to recently confirm their identification with almost full certainty. Just to give an idea of the time scale needed before an identification is widely accepted, the first claims of possible γ-ray emission from LSI+61◦ 303 were reported more than a quarter of a century ago[14]. From the theoretical point of view, at the time of writing this there is a strong debate about the origin of the TeV emission in γ-ray binaries, either in a microquasar jet or a binary pulsar scenario[20]. In the following sections, the contribution to γ-ray source identification being carried out by the author and his collaborators will be described along its Multi-wavelength Astronomy and the unidentified γ-ray sources 225 main lines. Our strategy is mainly based on intensive use of multi-wavelength data, both obtained by ourselves and from public archives. 4 Identifications based on variability criteria One of our lines of search consists on selecting the most variable EGRET sources in the vicinity of the galactic plane with |b| ≤ 10◦ . This choice is based on the variability index I, which establishes how variable is a given source as compared to the population of pulsars[25, 26]. Only sources with I ≥ 2.5 are included in this search, which are more than 3σ away from the statistical variations of pulsars. The sample of EGRET variables is further constrained by removing all cases where a plausible counterpart exists. Finally, we are left with a total of 10 objects to be explored. Both their variability and proximity to the galactic plane maximize the chances that these EGRET variables are galactic compact sources. Theoretical models of emission in γ-ray binaries and microquasars usually involve inverse Compton up-scattering of stellar optical/UV photons by relativistic electrons. These energetic particles are accelerated either in a jet scenario[7], or at the shock between stellar wind and the magnetosphere of a pulsar[12]. In both cases, γ, X-ray and radio emission is expected with some degree of variability in times scales of days to months (e.g. via orbital eccentricity). The finding of such a variable radio source, any of the radio emitters inside the EGRET contours, could easily betray the position of the EGRET counterpart. Based on these ideas, we are conducting a multi-epoch radio observation of the variable EGRET sources quoted above. Observations have been carried out using different interferometers, including the Very Large Array (VLA), the Westerbork Synthesis Radio Telescope (WSRT) and the Giant Metre Wave Radio Telescope (GMRT). The observations are being carried out at the 20-21 cm wavelength for easy comparison with the NRAO VLA Sky Survey (NVSS). The EGRET fields are covered as much as possible with mosaics of multiple pointings, distributed as illustrated in Fig. 3. This project is currently going on and nearly half of our EGRET fields have already been mapped[22] and a few variable radio sources found in each of them. In Fig. 4 we show the detection of one of such radio variables in the field of 3EG J1928+1733 by comparison with the NVSS. According to our criteria, this radio source becomes a candidate counterpart for follow up spectroscopic observations at optical and infrared wavelengths. It is only via spectroscopy that one can unveil the nature of each radio variable and assess the possible connection with its respective EGRET unidentified source. 5 Identifications based on ad-hoc observations Soon after a γ-ray or hard X-ray source is identified and an approximate position reported, it is sometimes possible to carry out ad-hoc observations or 226 J. Martı́-Ribas Galactic Latitude (deg) 3 2 1 18 17 16 Galactic Longitude (deg) 24 25 Galactic Longitude (deg) 26 15 23 Fig. 3. Two examples of our mosaicing strategy to cover most of the error boxes of variable EGRET sources. The circles represent the individual pointings of the GMRT at the 20 cm wavelength, where the antennae FWHM is about 30 arc-minute. The solid curves represent the 99, 95, 68 and 50% confidence contours of the EGRET position as in Fig. 1. Sub-image of the 3EG J1928+1733 field (21 cm) WSRT Sub-image of the 3EG J1928+1733 field (20 cm) 17 46 17 46 44 44 42 Declination (J2000.0) Declination (J2000.0) Galactic Latitude (deg) 4 40 38 NVSS 42 40 38 Variable 36 36 34 34 19 32 00 31 45 30 15 Right Ascension (J2000.0) Levels = 0.22 mJy/beam * (-3, 3, 5, 8, 10, 15, 20, 30, 40) 00 30 45 19 32 00 31 45 30 15 Right Ascension (J2000.0) 00 30 45 Levels = 0.4 mJy/beam * (-3, 3, 4, 5, 6, 8, 10, 12) Fig. 4. An example of a compact and variable radio sources discovered in the field of 3EG J1928+1733 when comparint a WSRT mosaic with the corresponding image of the NVSS at practically the same 20 cm wavelength. The proposed variable, indicated with a circle, appeared noticeably fainter a few years before our WSRT observations. Adapted from [22]. Multi-wavelength Astronomy and the unidentified γ-ray sources 227 inspections of survey data in order to quickly obtain a first idea of what the counterpart may be. This is an approach followed by many observers active in identifying the numerous sources discovered by the INTEGRAL satellite. Most of the relevant information is often spread in dedicated electronic publications6 well in advance of the final publication in professional journals. -14 54 52 6 cm Declination (J2000) Declination (J2000.0) 53 54 2MASS J1802473-145454 55 NVSS J180247-145451 56 57 18 02 47.6 Right Ascension (J2000.0) 47.5 47.4 47.3 Right Ascension (J2000) 47.2 47.1 Right Ascension (J2000.0) Fig. 5. Left. The 90% confidence error circle of IGR J18027−1455 and the two radio sources inside it appearing in the NVSS. Only one of them is also consistent in position with the smaller error circle of a ROSAT X-ray source, thus rendering it as the most likely counterpart of the INTEGRAL source. Right. The radio counterpart candidate observed at 6 cm with the VLA in A configuration in order to obtain a sub-arcsecond accurate position. Both panels are adapted from [9]. In Fig. 5 we report results of this kind carried out by our team on the INTEGRAL source IGR J18027−1455. Originally, two NVSS radio sources were found inside the INTEGRAL error circle by [8]. One of them was further consistent with the only ROSAT soft X-ray source within the circle. This fact promptly led us to conduct further observations with the VLA, Calar Alto and ESO telescopes at radio, optical and near infrared wavelengths, respectively. The radio emission resulted compact and with a non thermal spectral index α = −0.75 ± 0.02 (Sν ∝ ν α ). The infrared images and optical spectra (see Fig. 6) were consistent with an extended source with broad reshifted Hα emission lines at z = 0.034. This pointed to an Active Galactic Nucleus (AGN) of Seyfert 1 type as the optical/infrared counterpart of IGR J18027−1455[9], in agreement with independent work[18]. In addition, our results interestingly showed as well that this Seyfert is intrinsically bright at high energies both from the absolute point of view and when scaled to a normalized 6 cm luminosity. Its X-ray luminosity was finally compared to isotropic indicators and the object could be classified as Compton thin and AGN dominated. All these observational facts are relevant for the understanding of INTEGRAL extragalactic sources. 6 See for instance http://www.astronomerstelegram.org/ J. Martı́-Ribas Declination (J2000.0) 4 19 July 2004 3 Relative Flux 228 2 1 0 500 550 600 700 650 Wavelength (nm) 750 800 850 900 Right Ascension (J2000.0) Fig. 6. Left. Near infrared Ks-band image of the counterpart candidate to IGR J18027−1455 taken with the ESO New Technology Telescope (NTT) showing the existence of an extended object in good positional agreement with the VLA radio position. Right. Optical spectra of the proposed counterpart acquired with 2.2 m telescope at CAHA and the CAFOS spectro-imager. A prominent broad redshifted Hα line is detected that allowed us to classify it as a Seyfert 1 galaxy in agreement with independent work[18]. Both panels are also adapted from [9]. 6 Identifications based on modern radio surveys Another powerful tool to be used for γ-ray counterpart searches is the on-line availability of different radio and X-ray surveys and data archives. Fig. 7 is an example of such survey assisted identifications for the case of the soft γray source KS 1741−293. Originally discovered as a hard X-ray transient by the KVANT module of the Mir space station[27], its position has remained poorly accurate at the ∼ 10 for more than a decade despite been heavily scrutinized[10]. KS 1741−293 is today known to be among the high energy sources dominating the Galactic Center sky in hard X-ray/soft γ-rays ([6]) and its true nature still remains unknown. From its apparent bursting nature in the discovery years, a neutron star low-mass X-ray binary interpretation has been proposed. However, the absence of an optical, infrared or radio counterpart makes difficult to confidently test such hypothesis. Our interest about KS 1741−293 started recently from the identification of a possible radio counterpart using the Multi-Array Galactic Plane Imaging Survey (MAGPIS) at the 6 cm wavelength[16]. A very deep radio map of the field was then assembled by combining different observations from the VLA archive into the single map of Fig. 7. The image here shows a rich field with a non-thermal radio source reminiscent of a supernova remnant. Multi-wavelength Astronomy and the unidentified γ-ray sources 229 6 cm INTEGRAL IBIS/ISGRI -29 19 30 DECLINATION (J2000) 20 00 30 KVANT 21 00 Chandra 30 ASCA 22 00 30 17 44 58 56 54 52 50 48 RIGHT ASCENSION (J2000) 46 44 Fig. 7. Identification of a very likely candidate counterpart to the soft γ-ray source KS 1741−293, which was previously detected by KVANT, ASCA and INTEGRAL IBIS/ISGRI. The corresponding 90% confidence error circles are also shown and labelled accordingly. The proposed counterpart[17] is a Chandra X-ray source located with arc-second accuracy at the star symbol position. Contours trace the radio emission as observed with the VLA at the 6 cm wavelength. They correspond to −4, 4, 5, 8, 10, 12, 16, 20, 24, 28 and 32 times 0.4 mJy beam−1 , the radio rms noise. Interestingly, the Chandra X-ray image of the region reveals only a single compact X-ray source in the field consistent with all the 90% confidence circles of the different space missions having detected KS 1741−293 (KVANT, ASCA and INTEGRAL). This fact leads us to propose this Chandra X-ray source, namely CXOGC J174451.0−292116, as the likely counterpart of KS 1741−293. The Chandra data provide us with an arc-second accurate position where to search for the near infrared counterpart of the system in order to confirm the identification and the proposed low-mass X-ray binary nature. Being very close to the Galactic Center, we are dealing with a heavily absorbed object and preliminary searches using the 2 Micron All Sky Survey[24] and deep optical imaging with the Spanish 1.52 m telescope at Calar Alto (Spain) have failed to reveal any counterpart. However, both the infrared and optical magnitude limits so far obtained can still be significantly improved, for instance with a 4 m class telescope, and a deeper search is in progress. The reader is referred to [17] for a more detailed account of this on-going work. 7 General conclusions and future prospects Astrophysics in γ-rays is in a clear course to reaching maturity in the next few years. This spectral window has revealed a Universe full of unknown objects 230 J. Martı́-Ribas and their understanding will be clearly very rewarding from the Physics points of view. However, this task must overcome the technical difficulties of γ-ray imaging with both ground and space based observatories. The examples of identification discussed in this review provide an idea of the degree of difficulty involved. In any case, the present generation of Cherenkov arrays and the upcoming launch of GLAST and other missions anticipate excellent conditions for the development of this promising field. Acknowledgements: The author acknowledges support by grant AYA2004-07171C02-02 of Spanish government, FEDER funds, and FQM322 research group of Junta de Andalucı́a. References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. J. Albert, et al.: Science 312, 1771 (2006) F. Aharonian, et al.: A&A 393, L37 (2002) F. Aharonian, et al.: A&A 442, 1 (2005) F. Aharonian, et al.: A&A 439, 1013 (2005) W. Baade, R. Minkowski: ApJ 119, 206 (1954) G. Bélanger, et al.: ApJ 636, 275 (2006) V. Bosch-Ramon, et al.: A&A 447, 263 (2006) J. A. Combi, et al.: ATel 246 (2004) J. A. Combi, et al.: A&A 458, 761 (2006) A. M. Cherepashchuk, et al.: A&A 289, 419 (1994) A. E. Chudakov, et al.: J. Phys. Soc. Japan 17, A-III 106 (1962) G. Dubus: A&A 456, 801 (2006) S. Funk: in The Multi-Messenger Aproach to High-Energy Gamma-Ray Sources, J. M. Paredes, O. Reimer and D. F. Torres editors, in press (2007) P. C. Gregory, et al.: AJ 84, 1039 (1979) R. C. Hartman, et al.: ApJSS 123, 79 (1999) D. J. Helfand, et al.: AJ 131, 2525 (2006) J. Martı́, et al.: A&A in press (2006) N. Masetti, et al.: A&A 426, L41 (2004) N. Masetti, et al.: A&A 455, 11 (2006) I. F. Mirabel: Science 312, 1759 (2006) J. M. Paredes, et al.: Science 288, 2340 (2000) J. M. Paredes, et al.: Astrophys. and Space Sci. 297, 233 (2005) P. V. Ramana Murthy, A. W. Wolfendale: Gamma-ray astronomy, Cambridge University Press (1993) M. F. Skrutskie, et al.: AJ 131, 1163 (2006) D. F. Torrres, et al.: A&A 370, 468 (2001) D. F. Torrres, et al.: Astron. Nach. 322, 223 (2001) J. J. J. in’t Zand, et al.: Adv. Space Res. 11, No.8, 187 (1991) The disc and plane of the Milky Way in the Near Infared A. Cabrera-Lavers1,2, M. López-Corredoira1, F. Garzón1,3 , P.L. Hammersley1 , C. González-Fernández1, B. Vicente1 1 2 3 Instituto de Astrofı́sica de Canarias (IAC), C/Vı́a Láctea, s/n, E-38200, La Laguna, Tenerife, Spain, antonio.cabrera@gtc.iac.es GTC Project Office, C/Vı́a Láctea, s/n, E-38200, La Laguna, Tenerife, Spain. Departamento de Astrofı́sica, Universidad de La Laguna, S/C de Tenerife, Spain. Summary. Near Infrared (NIR) data constitute a very valuable tool for analyzing the structure of the Milky Way. In this wavelength the emission is dominated by K and M giants, and it is less affected of extinction than the optical bands. Therefore, it reflects better the true distribution of stars in the Galaxy with a higher penetration in the more obscured zones of the Milky Way, mainly those near the Galactic plane. In this contribution we present some of our lattest results in the analysis of the structure of the Galactic disc by means of near-plane data in the NIR. 1 NIR Galactic surveys and Galactic structure Star counts have long been used to examine the stellar contents in the Galaxy (see [12]). However there are still controversial or totally unknown parameters in the description of the detailed stellar structure, some of them concerned with the radial and vertical distribution of the galactic disc, and its specific morphology. The NIR absolute magnitude of most giants stars is sufficiently bright to allow them to be easily detected deep into the Galactic plane, whilst the range of intrinsic J −K (or J −H) colour is sufficient to separate them on a colour-magnitude diagram (CMD). As well as being able to detect sources to a greater distance, the infrared source counts are far less affected by local dust, and so the measured distribution of sources is closer to the true distribution. For this reason, in the past two decades there have been large advances in this topic with the combined use of detailed models of stellar distribution ([18]; [3]; [16]) along with large area, high sensitivity and multi-colour star counts surveys as for example: TMGS ([9]), DENIS ([7]), 2MASS ([17]) and more recently, TCS-CAIN ([4]), a private NIR survey developed at the IAC with nearly 42 deg2 of the sky observed simultaneously in the J, H and K bands more than 1.5 mags deeper than 2MASS in near-plane regions of the Galaxy. 232 A. Cabrera-Lavers et al. 2 The red clump method Appart of the well-known method of inverting the star counts, we have developed a method of deriving stellar densities and the interstellar extinction along a given line of sight. We use the population of red clump giants as they are by far the more prominent population of the disc giants ([5]; [11]) (see Fig. 1). Their mean absolute magnitude and intrinsic color are assumed to be MK =-1.65 and (J − K)0 =0.75, with a Gaussian dispersion of 0.3 mag in absolute magnitude and 0.2 in color ([8]). These values are in good agreement with recent results in open clusters ([1]; [10]), with small dispersion due to metallicity or age gradients. This makes the K band magnitude of the red clump a good distance indicator ([13]), allowing spatial information to be extracted from the CMDs. 0.9 1 0.8 Normalized density (star pc−3) 0.9 FL ( star pc −3 ) 0.7 0.6 0.5 0.4 0.3 0.2 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.1 0 0.8 0 −3 −2 −1 0 Absolute magnitude K 1 0.6 0.8 1 1.2 1.4 (J−K)0 Fig. 1. Distribution of K absolute magnitudes and (J − K)0 colors corresponding to the sum of all giants types in the disc according to the ”SKY” model. The maxima correspond to the red clump (MK =-1.65, (J − K)0 =0.75), which are predominant in the giant population. To apply the method, theoretical traces of different spectral types based on the updated ”SKY” model ([18]), are first used to define the K giant branch in the CMDs. The giant stars are extracted from the CMDs and binned in apparent K magnitude. For each magnitude bin, count histograms in color are constructed. A Gaussian function was then fit to the histograms to determine the color of the peak counts at each magnitude (see Fig. 2). The extinction AK (mK ) can be determined by tracing how the peak (J − K)mK of the red clump counts changes with mK , and the intrinsic mean color (J − K)0 of the red clump. From the color excess and after [14]: (J − K)mK − (J − K)0 (1) 1.52 A mean distance can be assigned given the mean absolute magnitude of the red clump giants, hence giving the interstellar extinction along the line AK (mK ) = The disc and plane of the Milky Way in the Near Infrared 233 l=65o b=0o 6 8 K 10 12 14 16 0 1 2 3 4 J−K Fig. 2. Left: Extracting the K giant stars. Dashed lines show the selected region isolating the red clump giants and filled circles show the maxima of the red clump for individual magnitude bins obtained via Gaussian fit. Right: Example for three magnitude bins in the field l=65◦ b=0◦ . l=15o 6 b=2o l=220o 6 8 10 10 K K 8 b=0o 12 12 14 14 16 0 1 2 J−K 3 4 16 0 1 2 J−K 3 4 Fig. 3. Extracting the red clump stars from the CMDs for two different lines of sight corresponding to the inner (left) and outer (right) Galactic disc. Solid line shows the fitted trace we assign to the red clump giants. The dashed lines show the limits for the red clump giants extraction within a width of 0.4 mag. of sight. Stellar density is obtained extracting the sources with a (J − K) within 0.2 mag of the center of the fitted red clump (see Fig.3). Extraction is limited to mK <13, to avoid any contamination in the star counts by the dwarf population. Once it is assumed that the red clump giants have a mean absolute magnitude of MK =-1.65, the luminosity function is replaced by a Delta function, giving directly the density as: D(r) = A(m)δm wr2 δr (2) 234 A. Cabrera-Lavers et al. with A(m) the number of stars per unit area of solid angle w at m in interval δm, and δr the related distance interval. 3 The outer disc We analyzed the old stellar population in the NIR data around the Galactic plane of the 2MASS survey. The disc is well fitted by an exponential distribution along both the galactocentric distance (R) and height (z): ρ(R, z) = ρ e− hz (R) = hz (R )e R−R hR,f lare R−R H |z| e− hz (R) (3) −1 −1 H = (h−1 R + hR,f lare ) (4) This stands for a density which falls exponencially with the galactocentric radius R (scalelength hR ), and with the heightness z (variable scalelength hz (R)). The scaleheight is variable due to the “flare”, which distributes the stars in a wider scaleheight which increases with galactocentric distance (we modeled this flare as an exponential increase of the scaleheight). We use two different methods applied to the 2MASS K-band star counts: isolating red clump giants in some few near plane regions to invert directly their star counts and obtain the disc density in several lines of sight (method A); and fitting accurately the disc model in 820 near plane regions (method B). The latter allows the contribution of the Galactic warp to be introduced. This warp bends the Galactic plane upwards at 0◦ <l<180◦ and donwards at 180◦ <l<360◦. To reproduce this, we fitted a model like the representation of eq. (3) but with |z − zw | instead of |z|, where zw , the elevation of the disc, is: zw = [Cw R(pc)w cos(φ − φw ) + 15] pc. (5) Table 1. Parameters extracted for the outer Galactic disc using the two methods applied to the 2MASS K-band star counts. selected areas H (kpc) hz (R ) (pc) hR,f lare (kpc) +0.22 l = 180 , 220 2.10−0.17 ◦ ◦ ◦ ◦ ◦ b = 0 , 3 , 6 , 9 , 12 +0.20 Method B 45◦ < l < 315◦ 1.90−0.16 ◦ ◦ ◦ ◦ without warp b = 0 , ±3 , ±6 , ±9 +0.15 Method B 45◦ < l < 315◦ 1.97−0.12 ◦ ◦ ◦ ◦ with warp b = 0 , ±3 , ±6 , ±9 Method A ◦ ◦ 310+60 −45 3.4±0.4 300+13 −15 4.6±0.5 285+8 −12 5.0±0.5 As Table 1 shown both methods give consistent results, thus this can be used as a test of the reability of the red clump method in deriving structural parameters of the disc. Summarizing the results from the fitting to the star The disc and plane of the Milky Way in the Near Infrared 235 counts with the inclusion of the warp (as they are the more precise ones) we obtained that the scaleheight in the solar disc is hz (R )=285+8 −12 pc, and the intrinsic scalelength (that is, without the effect of the flare) is hR =3.3+0.5 −0.4 kpc. There is a strong flare towards the outer Galaxy, that follows roughly a law hR,f lare =12-0.6 R (kpc) kpc (for R <15 kpc). Also, there is not apparently an abrut cut-off in the stellar disc (at least within R <15 kpc) in contradiction with previous estimates (e.g., [15]). The best fit to the warp model (w = 5.25±0.5; φw =-5◦ ±5◦ ) gives results consistent with other works (e.g., [6]) but it is interesting to note that the amplitude of this stellar warp is coincident with that of the gas. Whatever it is cause of the warp, it affects both to the stars and gas in the same rate. o b=0 −5 −5 10 10 ρ(R,z) ρ(R) R=10 kpc R=12 kpc R=14 Kpc R=10 kpc R=12 kpc R=14 Kpc −6 10 o l=180 o l=220 o l=155 o l=165 Model with H=2.1 kpc −6 10 −7 10 8000 10000 12000 R (pc) 14000 16000 0 500 1000 1500 z (pc) Fig. 4. Left: Fit of the star counts in the plane using method A. Right: Stars density obtained from the data and the lines standing for the model with H=2.1 kpc, hz (R )=310 pc, hR,f lare =3.4 kpc. Note that the data show a slower decrease of density for larger R, and this fact can be modeled with the flare. 4 The inner disc We applied the red clump method for deriving the stellar densities in 6 regions towards the inner Galaxy where the inner disc can be isolated from other components (1.5◦ <|b|<6.5◦, 15◦ <l<20◦), thus we avoid the in-plane regions where the bulge, spiral arms, stellar ring and/or the Galactic bar might contaminate the counts. We have obtained in this case that the scaleheight increases towards the centre with the opposite trend to that of the outer disc described in Sect. 3. A weighted fit of a linear law of hz gives: hz = 317 ± 17 − [R(kpc) − 4]48 ± 20 pc; for 2.25 kpc < R < 4.25 kpc (6) Once we know the scaleheight as a function of the radius, hz (R), we can calculate indirectly the in-plane density as follows: ρK2 III (R, z = 0) ≡ ρK2 III (R, z) −|z| e hz (R) , (7) 236 A. Cabrera-Lavers et al. where hz (R) is taken from eq. (6) for R < 5.9 kpc and from eq. (4) for R > 5.9 kpc, in order to keep the continuity of hz . This is represented in Fig. 5, where it can be observed that the counts in the plane (left panel) do not follow the predictions of the exponential model for the disc density of Sect. 3 with a noticeable deficit of stars in the innermost 4 kpc with respect to the extrapolation of the exponential law. However, this deficit of stars disappears for higher z (right panel in Fig. 5), so it is only affecting to regions near the plane. Fig. 5. Density of red clump giants in the plane (left) and at fixed z=400 pc (right), derived indirectly from the counts, using the hz (R) from eq. (6) for R < 5.9 kpc and from eq. (4) for R > 5.9 kpc. Solid line shows the extrapolation of the exponential model obtained in Sect. 3, while the rest of lines stand for other alternatives that fit even better the data, as a constant density for the inner 4 kpc, a modified exponential with a ’hole’ in the inner 4 kpc, or even a strong flare in the inner disc. We also compare the disc models with star counts both in the NIR, representative of the old disc population, and MIR (i.e., a young population). In this latter there is practically zero extinction, so this has nothing to do with the observed results. The comparison is carried out in Fig. 6 with the predictions of two exponential models: [8] (L02), based on star counts, and [3] (B02), based on flux maps. It is clearly seen that counts are nearly constant between |l| = 15◦ and |l| = 30◦ and this cannot be fitted with a purely exponential disc in the inner stellar disc. Therefore, there is a deficit in the distribution of stars near the plane in the inner 4 kpc of the Milky Way respect to the predictions of a pure exponential disc model. This deficit is present both in the young and old populations thus is probably a rather stable feature of the disc which might due to the existence of an in-plane bar that sweeps out the near-plane stars. The disc and plane of the Milky Way in the Near Infrared o mK<9.0, −2.0 <b<−1.5 −0.5 <b<0.5 MSX B02 disc model L02 disc model L02 with hole 200 3000 −2 N (deg ) −2 237 o 300 DeNIS TMGS L02 with hole L02 without hole B02 disc model 4000 Star counts (deg ) o o 2000 100 1000 0 40 20 0 −20 Galactic longitude (deg.) −40 0 −45 −30 −15 0 l (deg.) 15 30 45 Fig. 6. Left: DENIS/TMGS star counts with mK ≤ 9.0 for −2◦ < b < −1.5◦ . Right: MSX star counts in the plane at 14.6 µm up to magnitude 3.0. In both cases, comparisons with the B02 disc model and the extrapolation of the L02 model of outer disc towards the centre with either an exponential (’L02 without hole’) or a flat density distribution (’L02 with hole’) are also shown. Note that the models represent only the disc (without the bulge, bar and ring) while the counts do include everything in the line of sight. 5 A smooth model for outer and inner disc With the results summarized in Sect. 3 and 4 we can construct the following picture for the disc of the Milky Way: We have obtained that the inner disc presents a deficit of stars respect to the predictions of a pure exponential for the innermost 4 kpc, where the in-plane density remains nearly constant. The outer disc is well repoduced by a double exponential law and shows a strong flare that moves the stars to higher heights above the plane as we move outwars in the Galaxy. In this outermost regions is also where the assymetries due to the Galactic warp are higher, displacing the mean disc (z=0) up to 2 kpc between the location of the maximun and the minimum amplitude of the warp. Numerically, an expression that summarizes the outer disc (R > 6 kpc) and the inner disc (2.5 kpc < R < 4 kpc) with a smooth transition between two regimes is: R 3740 pc R + 3740 pc ρ(R, z) ≈ ρ e 1970 pc R e−( 1970 pc + R ) e−|z|/hz , (8) hz ≈ 285[1 + 0.21 kpc−1 (R − R ) + 0.056 kpc−2 (R − R )2 ] pc. (9) This density is shown in Fig. 7. As it can be noted, the disc of the Milky Way is far from being as smooth and well-behaved as it was assumed in the pioneering works of modelling the distribution of stars in the Galaxy, when a simple symmetric exponential was enough to reproduce the observed counts (e.g., [2]). Nowadays, we deal with the more complete star counts databases ever, and the more sophisticated and complex Galactic models to date are 238 A. Cabrera-Lavers et al. currently being constructed. Therefore, we are now in the right way for really unveiling the real structure of Our Galaxy. 1.0 z(Kpc) 0.5 0.0 -0.5 -1.0 -10 -5 0 R(Kpc) 5 10 Fig. 7. Contour diagram of log10 ρ (pc−3 ) of a possible interpolated/extrapolated smooth model of the disc according to eq. (8) for R < R and eq. (3) for R ≥ R in the yz-plane of the Galaxy (perpendicular to the line Sun–Galactic Centre) with y between −12.0 and 12.0 kpc and z between −1.0 and 1.0 kpc (vertical scale in the plot multiplied by a factor 5). Lower contour: log10 ρ (pc−3 ) = −2.1; step = 0.15. References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. Alves, D.R., ApJ, 539, 732 (2000) Bahcall, J.N. & Soneira, R.M., ApJS, 44, 73 (1980) Bissantz, N., & Gerhard, O.E., MNRAS; 330, 591 (2002) Cabrera-Lavers, A., Garzón, F., Hammersley, P.L., et al., A&A, 453, 371 (2006) Cohen, M., Hammersley, P.L., & Egan, M.P., AJ, 120, 3362 (2000) Drimmel, R., & Spergel, D.N., ApJ, 556, 181 (2001) Epchtein, N., Deul, E., Derriere, S., et al., A&A, 349, 236 (1999) López-Corredoira, M., Cabrera-Lavers, A., Garzón, F. & Hammersley, P.L., A&A, 394, 883 (2002) Garzón, F., Hammersley, P.L., Mahoney, T., et al., MNRAS, 264, 773 (1993) Grocholski, A.J., & Sarajedini, A., AJ, 123, 1603 (2002) Hammersley, P.L., Garzón, F., Mahoney, T.J., et al., MNRAS, 317, L45 (2000) Paul, E.R., in The Milky Way Galaxy and Statistical Cosmology, 1890-1924 (Cambridge University Press, Cambridge) (1993) Pietrzyńsky, G., Gieren, W., & Udalski, A., AJ, 125, 2494 (2003) Rieke G. H., & Lebofsky M. J., ApJ 288, 618 (1985) Robin, A., & Crézé, M., A&A, 157, 71 (1986) Robin, A., Reylé, C., Derriére, S., & Picaud, S., A&A, 409, 523 (2003) Skrutskie M. F., Cutri, R.M., Stiening R., et al., AJ, 131, 1163 (2006) Wainscoat, R.J., Cohen, M., Volk, K., et al., ApJS, 83, 111 (1992) AGB Stars: Nucleosynthesis and Open Problems I. Domı́nguez1 , C. Abia1 , S. Cristallo2 , P. de Laverny3, and O. Straniero2 1 2 3 Dpto. Fı́sica Teórica y del Cosmos, Universidad de Granada, 18071 Granada, Spain, inma@ugr.es INAF-Osservatorio Astronomico di Teramo, Via M.Maggini 47, 64100 Teramo, Italy Observatoire de la Côte d’Azur, Dpt. Cassiopee UMR6202, 06304 Nice Cedex 4, France Summary. AGB stars play a fundamental role in the chemical evolution of the Universe, being the main contributors to many elements beyond the Fe-peak, as well as important producers of 7 Li, 12 C, 14 N, 26 Al and other relevant isotopes. These elements are synthesized in their interiors through complex and challenging physical processes, transported to the surface and injected into the interstellar medium via mass loss. Moreover, in their extended envelopes stellar grains of different types are formed from whose chemical analysis a valuable information can be extracted about the nucleosynthetic processes occuring in these stars. In the last years there has been a significant improvement in the field due to the refinement of the numerical models and to the new high resolution optical and infrared spectroscopic studies, including the first extragalactic AGB stars. In addition, the chemical abundances derived in some extremely metal poor stars and in the intergalactic medium at high redhift could be interpreted as being produced by the first population, zero metals, AGB stars. In this talk I will review the state of the art in AGB modelling with special emphasis on nucleosynthesis. 1 Introduction Most stars become Asymptotic Giant Branch (AGB) stars at the end of their evolution. At the AGB phase the extended convective envelope penetrates inward to the zones already processed by nuclear burning and the mass loss rate increases. About 75% of all the mass returned by stars to the interstellar medium (ISM) comes from AGB stars [11, 23, 21]. The AGB phase ends when the envelope is ejected and a Planetary Nebulae is formed. The central star is a white dwarf (WD) which in case of binary evolution could be the progenitor of a novae or a thermonuclear supernova. AGB stars have initial masses smaller than 8 M , the exact value depends on metallicity (and on the different numerical codes). Stars with higher masses 240 I. Domı́nguez et al. develop a carbon and oxygen (CO) core massive enough (≥1.1 M ) to reach the temperature needed for carbon ignition. These massive stars would likely explode as core collapse supernovae, injecting to the ISM nuclei produced during the advanced nuclear phases. The internal structure of an AGB star consists of an inert, partially degenerate CO core, a He shell separated from the H shell by a He-rich region (He intershell) and an extended convective envelope. The luminosity of the star is mainly provided by the H burning shell, located just bellow the bottom of the convective envelope. This burning is repeatedly interrupted by the He shell burning, which ignites in runaway conditions. This phenomenon is known as a thermal pulse (TP). These TPs cause an expansion of the overlying layers and the H shell is temporaly extinguished (or nearly extinguished). Withouth the activity of the H shell the convective envelope penetrates inward, in the C-rich He intershell, and carries to the surface the elements produced in this zone (Third Dredged Up episode, TDU). This is the origin of the numerous chemical and spectroscopic peculiarities observed in AGB stars (intrinsic) or in stars that have accreted matter from an AGB companion in a binary system (extrinsic). Many AGB stars have a C/O ratio (by number) ≥1 in the envelope, which is the definition of a carbon star (C-star). Since the majority of stars are born with a C/O ratio≤1 (the solar value is C/O∼0.54), this carbon enrichment must result from a deep mixing process (the TDU described above). It is believed that about 30% (maybe even more) of the observed carbon is produced by AGB stars, as well as an important contribution to nitrogen. Both elements are crucial for organic chemistry and life cycles. It is also remarkable the role played by AGB stars in the 7 Li nucleosynthesis, crucial for understanding the evolution of Li abundance in the Galaxy. Beyond the Fe peak (A≥56), the Coulomb barrier is too high and the synthesis of heavier nuclei is produced by neutron captures. When the characteristic time for neutron capture is shorter than β decay lifetimes, the process is known as r-process (rapid) and as s-process (slow) when it is larger . The r-process requires high neutron densities (≥1020 cm−3 ) and it is believed to occur in core collapse supernovae, although the identification of a satisfactory astrophysical site is still a matter of debate. The s-process operates at densities about 10 order of magnitudes smaller. These neutron densities are found in the He shell and He intershell during the AGB phase. In AGB stars, neutrons come from two reactions: 13 C(α,n)16 O that operates at temperatures of ∼90 106 K and produces low neutron densities (∼107 cm−3 ), and 22 Ne(α,n)25 Mg that operates at higher temperatures (∼ 300 106 K) and produces higher neutron densities (1011 cm−3 ). The difference in temperature relates these two neutron sources to a different range of stellar masses. Note that 22 Ne is produced during the initial phase of He burning (all 14 N is converted to 22 Ne), and hence is naturally present in the He convective shell. This is not the case for 13 C. It is believed that 13 C is AGB Stars: Nucleosynthesis and Open Problems 241 produced during the interpulse period, in the He intershell, by proton capture over the abundant 12 C. This requires mixing of protons from above. 2 The Stellar Evolutionary Models The models presented in this talk have been obtained with the FRANEC code ([13] and references therein), following the evolution of the star from the pre-main sequence to the end of the TP-AGB phase. In the present version of the code [17], the physical and chemical evolutions are fully coupled and an extended nuclear network with more than 750 reactions and 500 isotopes, from H to Pb/Bi, has been included. The opacity is calculated taken into account the changes in the internal chemical composition by interpolating linearly between tables with different Z and Y. This method is accurate enough as far as the relative distribution of the elements does not change. Unfortunately, this is not the case of low metallicity AGB stars where the metal increase in the envelope is mainly due to the 12 C dredged up. However, up to date no low temperature opacity tables with carbon and/or nitrogen enhancements are available and, therefore we tentatively simulate the envelope enrichment by interpolating the opacity coefficients in metallicity. The mass loss rate has been derived as a function of the pulsational period taken into account the rates obtained observationally for a sample of Galactic C and O-rich AGB stars. Note that the mass loss rate determines the duration of the AGB phase: the number of TPs and TDU episodes and thus, influences the final chemical yields. We have introduced a physical algorithm for the treatment of the convective/radiative interfaces [13]. This leads to the diffussion of protons from the H shell to the He intershell and produces 13 C, the neutron source, by the reaction 12 C(p,γ)13 N(β + ν)13 C. Neutrons are captured by Fe-seeds (or lighter isotopes in low metallicity stars) and the s-process path is opened. 3 Galactic Carbon C(N) Stars A 2 M solar metallicity model has been computed as representative of the AGB disk population of our Galaxy (for details see [17]). We calibrate a standar solar model taken into account the new solar abundance determination of C, N and O [4, 6] and diffussion processes. We obtain for the initial solar values of Z and Y, 0.015 and 0.27, respectively. The star becomes a C-star after six TDU episodes and, at the end of the evolution (after 11 TDU episodes), the C/O ratio is about 2. The surface abundances of all heavy elements (from Sr to Pb) are enhanced. We find that the abundance of the s-element first peak: Sr, Y and Zr (the so called low-s or ”ls” elements) are comparable to those of the second peak: Ba, La, Ce, Pr 242 I. Domı́nguez et al. and Nd (high-s or ”hs” elements), [ls/Fe]=+1.1 and [hs/Fe]=+0.81. The third peak (Pb) is underproduced with respect to the second (Ba) as expected for solar metallicity, [Pb/Fe]=+0.5. These results are compared with our observed sample of Galactic carbon C(N) stars [2, 3] for which, based on high-resolution spectra, we deduce a C/O∼1, and, on average, [ls/Fe]=+0.7±0.1 and [hs/Fe]=+0.5±0.3. When C/O in the envelope is C/O=1, our model shows [ls/Fe]=+0.8, [hs/Fe]=+0.55 and [hs/ls]=+0.7 in good agreement with the observational data. 3.1 The Mass of C(N) Stars The abundance ratios obtained, from high-resolution spectra, of Rb, Sr, Y, and Zr in our sample of C(N) type stars, allow us to get information on the neutron density during the neutron capture nucleosynthesis processes occurring in the He intershell. The critical reaction branching at 85 Kr determines the relative enrichment of the studied species, Rb abundance can differ by 1 order of magnitude depending on the neutron density. Therefore, the relative abundance of Rb to other elements in this region of the s-process path, such as Sr, Y, and Zr, can be used to estimate the average neutron density of the s-process. As quoted in the introduction, in AGB stars, low and high neutron densities are related to different neutron sources, 13 C and 22 Ne, respectively, operating at a different range of temperature and hence, at a different range of stellar masses. We have studied the 85 Kr branching on the s-process path and compared the observed abundance ratios to the model predictions for low and intermediate mass stars [2]. The main implication of such analysis is that most C(N) stars experience s-process nucleosynthesis phenomena dominated by the neutron source provided by α-captures on 13 C in radiative conditions, concluding that the majority of C(N) stars are of low mass (≤3 M ). 3.2 Early Solar System Radioactivity The presence of short-lived (meanlifes≤10 Myr) radioactive nuclei in the Early Solar System (ESS) has been established from measurements of their decay products on primitive meteoritic samples [24]. It was early proposed that these nuclei come from long-term galactic nucleosynthesis. This hypothesis works for some of them, but the ISM equilibrium abundances of 26 Al, 41 Ca and 60 Fe are too low, and a late injection to the protosolar nebula is required. A nearby core collapse SN could be responsible for most of the short-lived nuclei but it overproduces 53 Mn. Wasserburg [30] proposed that a low-mass AGB star could be the source of 41 Ca, 60 Fe, 107 Pd and 26 Al [32]. All the above nuclei are produced during the TP-AGB phase in our 2 M solar metallicity model. However, when we adjust the dilution factor (the ratio of the contaminating mass to the solar nebula mass) to reproduce 107 Pd, the AGB Stars: Nucleosynthesis and Open Problems 243 others, 26 Al, 41 Ca and 60 Fe, are underproduced. The situation is much better at the second TDU, this suggests that the polluting source was a lower mass AGB star (less TDU episodes), of about 1.3 or 1.5 M [18]. We are currently working on this topic. 4 Extremely Metal Poor C-rich Stars Spectroscopic surveys for very-metal poor ([Fe/H]≤-2.5) stars show that 20% to 30% of the candidates are carbon rich [8, 14, 9]. These halo stars are of such low mass that the TDU never occurs [27]. It is believed that their carbon enrichment is due to a previous accretion process from a companion, an AGB star (now a WD). If this is the case, we expect to observe s-process elements in these metal poor C-rich stars. High resolution spectroscopy confirms this picture [5, 7, 15, 29]. We have calculated a low metallicity model (2 M , [Fe/H]=-2.2), which may be considered as representative of metal poor AGB stars. Note that the abundance of 13 C (the main neutron source) in the He intershell does not depend on the initial metallicity. Thus, more neutrons per seed are available as the initial metallicity decreases and, after a few TDU episodes, the star becomes enhanced in Pb. At the end of the AGB phase, our model shows [C/Fe]=+3.3, [ls/Fe]=+1.7, [hs/Fe]=+2.3 and [Pb/Fe]=+3.1. These values are in good agreement with those measured in halo stars of similar metallicity [5, 7, 15, 29]. However, there is a problem related with nitrogen. Most of the observed metal poor s-rich stars are also N-rich, while in our model [N/Fe]=+0.6. Nitrogen may be produced in intermediate mass AGB stars at the bottom of the convective envelope (by means of the hot bottom burning process [26]) or in low mass AGB stars, by an extra-mixing taking place bellow the inner border of the convective envelope (known as cool bottom process, CBP [25]). Finally, in low mass models at very low metallicities ([Fe/H]≤-2.5), proton ingestion from the envelope into the He-convective shell can produce 14 N at the beginning of the AGB phase [22, 28]. 5 Extragalactic Carbon Stars With the aim of studying the dependence of the s-process (and mixing mechanism) occurring in AGB stars on stellar metallicity, we are performing a full chemical study of C-stars in the Local Group of galaxies: Magellanic Clouds, Sagittarius, Draco, Ursa Minor, Carina and Sculptor. These satellite galaxies span an interesting range of metallicities, -3≤[Fe/H]≤0 and some of them have experienced recent star formation episodes, which opens the possibility of observing intrinsic low metallicity C-stars. We have completed the analysis of three C-stars, one in the SMC and two in Sgr [19]. The abundance ratios derived between elements belonging to the 244 I. Domı́nguez et al. first and second s-process peaks agree remarkably well with the theoretical predictions of low mass metal-poor AGB models, in which the 13 C is the main source of neutrons. However, the observed C/O and 12 C/13 C ratios do not agree with theoretical predictions, requiring deeper dredge-up episodes or a non standard mixing process. 5.1 A Li-Rich Carbon Star in Draco: evidences of recent star formation and extra-mixing We have observed the first Li-rich star (log (Li)∼3.5), D461, in a metal-poor stellar population: the Draco dwarf spheroidal galaxy [20]. All the observed features (abundance pattern, gravity and effective temperature) are explained by our numerical simulations of a low mass (1.5 M ), low metallicity ([Fe/H]= -2) AGB star, undergoing TDU. However, to explain the observed Li, a downward extra-mixing (CBP type) is needed. In this way higher temperatures (large enough to produce 7 Be) are reached. The mixing must be fast enough to remove the fresh 7 Be from the hot bottom layers before it decays into 7 Li, which is easily destroyed (at temperatures of 2 106 K). Although the bulk of stellar population in Draco is dominated by very old stars (∼10 Gyr), the minimum mass for the occurrence of the TDU cannot be smaller than ∼1.3 M , which implies that D461 is younger than ∼3 Gyr and a recent star formation episode has occurred in Draco. 6 The First AGB Stars and the Chemical Evolution of the Universe Recent investigations of the chemical abundances in high-redshift systems (like the Lyman-α forest) and in extremely metal-poor stars in our Galaxy, allow us to search for the imprints of the first stars in these systems [1]. In order to do that we assume initial mass functions (IMF) proposed for the population III and study the consequence of this in the framework of a star burst and in a simple chemical evolution model. In these models we include the chemical yields obtained for strictly zero-metals stars in the mass range 3 to 200 M . Previously, we computed the evolution (and chemical yields) of zero-metal AGB stars [13] and we found, contrary to previous investigations, that these stars experience a normal TP-AGB phase with TDU episodes and, as a consequence, contribute significantly to the metal-enrichment of the ISM. Our models show that the large C and N enhancements [C,N/Fe]≥0.5 found in a significant fraction of extremely metal-poor stars in our Galaxy favor an IMF peaked at intermediate-mass stars. The metallicity observed in high-redshifts systems is easily reached for any of the IMFs adopted: stars are very efficient metal-producers. This limits the pregalactic star formation efficiency and hence, the contribution of the firts stellar remnants to the dark barionic matter in the Universe to less than 0.1%. AGB Stars: Nucleosynthesis and Open Problems 245 The available observational data do not allow to constraint further the IMF of the first stars. This situation is expected to change in the near future, as the high-redshift Universe is one of the main targets for large ground based telescopes and space missions. 7 Final Remarks The AGB phase is a fascinating field in which several areas of studies converged; from numerical simulations and high resolution spectroscopy (optical and IR) to experimental nuclear physics and laboratory analysis of the isotopic composition of meteorites. The abundance patterns found in galactic and extragalactic AGB C-stars, as well as the isotopic composition of presolar grains, can be explained by low mass AGB stars in which the neutron source is 13 C(α,n)16 O. This scenario can also explain the dependence of the s-process on metallicity observed in C-rich halo stars and extragalactic C-stars. 13 C is produced through the diffussion of protons from the H shell to the He intershell. This diffussion mechanism has still to be parametrized in the numerical simulations. There are still problems to reproduce the observed low C/O and 12 C/13 C in metal poor AGB stars enhanced in s-elements. It is possible that carbon condensates in grains or that dust formation obscures this phase. On the other hand, models do not explain the high [N/Fe] observed in extremely metal poor C-rich stars in the halo. The ingestion of protons from the envelope into the He-convective shell followed by a huge dredge-up, a mechanism already found in low mass AGB models at very low metallicities, may explain 14 N enrichments. Several hints indicate that some kind of extra-mixing (CBP) is needed: Li-rich AGB C-stars, low values of 12 C/13 C in RGB and C-rich AGB stars, and the isotopic ratios, 18 O/16 O, 17 O/16 O and 26 Al/27 Al, in meteorites [10, 25, 31, 33]. The physics behind this extra-mixing is still unknown; rotation or magnetic tubes [12] have been proposed. AGB modelling has significantly improved during the last years. However, we are still far from a proper treatment of the mixing regions: 3D hydrodynamic simulations are probably needed to treat consistently convection, TDU and CBP. Mass loss is a major unknown and it has a huge effect on the chemical yields: it determines the duration of the AGB and so, the number of TPs and TDU episodes. For solar metallicity, we rely on observations, for lower metallicities, on speculations. All these uncertainties are treated through parameters but AGB numerical simulations take a lot of CPU time (e.g. one model, one month) and this limits our exploratory power. 246 I. Domı́nguez et al. References 1. C. Abia, I. Domı́nguez, O. Straniero, M. Limongi, A. Chieffi: ApJ, 557, 126 (2001) 2. C. Abia, M. Busso, R. Gallino, I. Domı́nguez, O. Straniero, J. Isern: ApJ, 559, 1117 (2001) 3. C. Abia, I. Domı́nguez, R. Gallino, M. Busso et al.: ApJ, 579, 817 (2002) 4. C. Allende Prieto, D.L. Lambert, M. Asplund: APJ, 573, L137 (2002) 5. W. Aoki et al: ApJ, 580, 1149 (2002) 6. M. Asplund, N. Grevesse, A.J. Sauval, C. Allende Prieto, D. Kiselman: AA, 417, 751 (2004) 7. B. Barbuy et al: AA, 429, 1031 (2005) 8. T. Beers, G.W. Preston, S.A. Shectman: AJ, 103, 1987 (1992) 9. T. Beers, N. Christlieb: ARAA, 43, 531 (2005) 10. A.I. Boothroyd, I.J. Sackmann, G. Wasserburg: ApJ 430, 77 (1994) 11. M. Busso, R. Gallino, G. Wasserburg: ARAA, 37, 329 (1999) 12. M. Busso, A. Calandra, M.C. Nucci: MemSAIt, 77, 798 (2006) 13. A. Chieffi, I. Domı́nguez, M. Limongi, O. Straniero: ApJ, 554, 1159 (2001) 14. N. Christlieb, P.J. Green, L. Wisotzki, D. Reimers: 2001, AA, 375, 366 (2001) 15. J.G. Cohen , N. Christlieb, Y.Z. Quian, G.J. Wasserburg: ApJ, 588, 1082 (2003) 16. S. Cristallo, O. Straniero, R. Gallino, L. Piersanti, I. Domı́nguez: In From Lithium to Uranium: Elemental Tracers of Early Cosmic Evolution, ed by V. Hill, P. Francois, F. Primas (F. Cambridge: Cambridge University Press, 2005) pp.483–484 17. S. Cristallo: S-process nucleosynthesis in low mass AGB stars at different metallicities. PhD Thesis, Teramo University, Italy, available at http://www.oateramo.inaf.it/osservatorio/personale/cristallo (2006) 18. S. Cristallo, R. Gallino, O. Straniero, L. Piersanti, I. Domı́nguez: MemSAIt, 77, 774 (2006) 19. P. de Laverny et al.: AA, 446, 1107 (2006) 20. I. Domı́nguez, C. Abia, O. Straniero, S. Cristallo, Y. Pavlenko: AA, 422, 1045 (2004) 21. F. Herwig: ARAA, 43, 435 (2005) 22. D. Hollowell, I.Jr. Iben, M.Y. Fujimoto: ApJ, 351, 245 (1999) 23. I. Jr. Iben, A. Renzini: ARAA, 21, 271 (1983) 24. K.D. McKeegan, A.M. Davis: In Treatise on Geochemistry, (Elsevier Pergamon, Oxford, 2004), pp. 431 25. K.M. Nollett, M. Busso, G.J. Wasserburg: ApJ, 582, 1036 (2003) 26. I.J. Sackmann, A.I. Boothroyd: ApJ 366, 529 (1991) 27. O. Straniero, I. Domı́nguez, S. Cristallo, R. Gallino: PASA, 20, 289 (2003) 28. O. Straniero, S. Cristallo, R. Gallino, I. Domı́nguez: MemSAIt, 75, 665 (2004) 29. S. Van Eck, S. Goriely, A. Jorissen, B. Plez: AA, 404, 291 (2003) 30. G.J. Wasserburg, M. Busso, R. Gallino, C.M. Raiteri: ApJ, 424, 412 (1994) 31. G.J. Wasserburg, A.I. Boothroyd, I.J. Sackmann: ApJ, 447, L37 (1995) 32. G.J. Wasserburg, M. Busso, R. Gallino, K.M. Nollett: Nucl. Phys. A., 777, 5 (2006) 33. E. Zinner: AREPS, 26, 147 (1998) Studying galaxy formation and evolution from Local Group galaxies. C. Gallart Instituto de Astrofı́sica de Canarias, Spain, carme@iac.es Summary. In this contribution I present the main research activities of the IAC “Stellar Populations in Galaxies” research Group, with emphasis on the subtopics directly related with the study of the evolution of nearby galaxies. In particular, I discuss preliminary results of ongoing research on the Magellanic Clouds using deep ground-based observations, and on a sample of isolated Local Group dwarf galaxies, using data from the ACS on board the HST. Future plans with the GTC are discussed. 1 Introduction The process of galaxy formation and evolution is driven by two sets of parallel mechanisms: stellar formation, evolution and death, which drive the evolution of the stellar populations and of the gas and metal content of the galaxy, and the mass assembly process, which determines its morphological type and dynamical evolution, and which in turn may induce star formation. In the nearest objects, and in particular in those that can be resolved into stars like Local Group galaxies, these mechanisms can be studied in great detail. We are conducting a comprehensive study of Local Group galaxies using a number of complementary tools to shed light on these two main mechanisms that determine galaxy formation and evolution, namely: 1. The star formation history (SFH), and its influence on galaxy evolution, through: i) The study of deep colour–magnitude diagrams (CMDs) of each galaxy; ii) the spectroscopic abundances of resolved stars; iii) the analysis of the properties of their variable stars; iv) the study of the Milky Way Cluster system. 2. The mass assembly and the dynamical evolution of each system, through: i) The stellar population gradients and kinematics of stars of different ages; ii) the dynamics of the Local Group and the influence of interactions on galaxy evolution. 248 C. Gallart This project is designed to make the best use of the GTC and its first-light instruments, especially OSIRIS. In this contribution, I will present examples of ongoing research in several of the aspects quoted above. 2 The star formation history from deep colour–magnitude diagrams. The CMD, and in particular those reaching the oldest main-sequence turnoffs, is the best tool for retrieving in detail the SFH of a stellar system. In this case, information on the distribution of ages and metallicities of the stars present in the galaxy can be obtained directly from stars on the main sequence, which is the best understood phase of stellar evolution from the theoretical point of view. It is also the one in which stars are more separated in colour and magnitude as a function of age. The range of ages and metallicities present can be determined through comparison with theoretical isochrones. To quantitatively determine the SFH, it is necessary to compare the observed density distribution of stars with that predicted by stellar evolution models (see [11]). Our group has been traditionally dedicated to the derivation of SFHs from deep CMDs, through comparison of observed and synthetic CMDs (e.g. [9], [2], [5], [3]; see also http://iac-star.iac.es/iac-star/). We are currently involved in two major programmes in this regard. The first aims at providing spatially resolved SFHs for both Magellanic Clouds, using ground-based observations. These, however, have produced CMDs of quality comparable (but covering a much larger area) to CMDs obtained in the central regions of both objects using the WFPC2 on HST (e.g. [14], [17]). The second is devoted to obtaining detailed SFHs for a sample of isolated Local Group dwarf galaxies, using ACS CMDs reaching the oldest main-sequence turnoffs. 2.1 The LMC In the case of the LMC (and as part of the PhD thesis of I. Meschin), we have observed 12 half degree fields, six to the north and three each to the east and west of the galaxy centre, using the MOSAIC camera at the 4 m CTIO and the WFI at the 2.2 m in La Silla. These fields sample galactocentric distances (from 3◦ to 10◦ , or 2.6 to 8.8 kpc) not explored before to this photometric depth, i.e. with CMDs reaching the oldest main-sequence turnoffs. Through comparison with synthetic CMDs, these data will allow us to obtain detailed SFHs for all these fields and characterize the population gradients present in the galaxy. Figure 1 shows the CMDs for four fields observed to the north of the galaxy, obtained from the MOSAIC data. Isochrones from [16] in a suitable range of age and metallicity have been superimposed. It can be noticed that the age of the youngest population varies from field to field, in the sense that star formation has proceeded down to more recent epochs toward the galaxy centre: while in the fields situated at 3◦ and 5◦ , star formation has basically Studying galaxy formation and evolution from Local Group galaxies. 249 Fig. 1. CMDs for the four LMC fields observed with the MOSAIC Camera at the 4 m CTIO, located at ' 3◦ , 5◦ , 6◦ and 8◦ from the LMC centre (clockwise from top left, respectively). Isochrones from [16] (scaled solar, overshooting set), with the ages and metallicities indicated in the labels, have been superimposed. The locus of the zero-age horizontal-branch is that of the lowest metallicity considered. A distance modulus of (m − M )0 = 18.5 and reddenings of E(B − V ) = 0.1, 0.05, 0.04 and 0.03 have been assumed to transform the data to absolute magnitudes and colours. Determinations of the SFH of each field are under way through comparison with synthetic CMDs. continued to the present time, the field at 6◦ seems not to have stars younger than ' 200 Myr. Finally, the field at 8◦ formed the bulk of its stars before 2.5 Gyr, with some residual star formation up to 1.5 Gyr ago. The presence of an important intermediate-age population in this field, together with the fact that the surface brightness profile of the LMC remains exponential to this large galactocentric radius and shows no sign of disk truncation, led [10] to conclude that the LMC disc extends (and dominates over a possible stellar halo) out to a radius of at least 7 kpc. 250 C. Gallart 2.2 The SMC For the SMC (and as part of the PhD thesis of N. Noël), we have similar CMDs for 13 smaller fields observed with the 10000 telescope at LCO ([15]). These fields are distributed in different parts of the SMC such as the “Wing” area and to the west and south, and in a range of galactocentric distances (from ' 1◦ to 4◦ , or 1 to 4.1 kpc). Several studies (e.g. [19], [6]) have found that the SMC intermediate-age and old population has a spheroidal distribution, and that the asymmetric appearance of the SMC is primarily caused by the distribution of young stars. With our deeper data, we can shed new light on the age distribution of these structures. In particular, we confirmed that the underlying spheroidal population is composed of both intermediateage and old stars, and found that its age composition does not show strong galactocentric gradients. The three fields situated in the “Wing” region show very active current star formation, but only the one closer to the centre seems to present a substantial enhancement in recent star formation with respect to a constant SFR(t). The fields corresponding to the western side of the SMC present a much less populated young main sequence as compared with those on the east side, even at similar galactocentric radius, with signs of a greatly diminished SFR(t) from 2 Gyr ago to the present time. As in our LMC study, none of the studied fields, out to a galactocentric radius of 4◦ (or 4.2 kpc), is dominated by an old stellar population. 2.3 Isolated Local Group dwarf galaxies We are participating in two HST-ACS programs (P. ID: 10505, P.I. Gallart; P.ID: 10590, P.I. Cole) with a total of 113 awarded orbits, in addition to an HST-WFPC2 program (P.ID: 8706, P.I. A. Aparicio), to obtain CMDs reaching the oldest main-sequence turnoffs in 6 isolated Local Group galaxies (two dIrr galaxies: Leo A and IC1613, the two isolated dSph galaxies discovered so far in the Local Group: Cetus and Tucana, and two transition type dIrr/dSph galaxies: LGS3 and Phoenix). Figure 2 shows the CMDs of four of the galaxies in the sample. Note the variety of SFHs, as hinted at by the comparison with selected isochrones from [16]. This is the first time that data of this high quality has been obtained for dwarf galaxies beyond the Milky Way satellite system. These data will allow us to obtain detailed and accurate SFHs for all these systems, through comparison with synthetic CMDs, and using additional constraints from the characteristics of their variable star population (see Section 4 below, and the contribution by E. Bernard et al. in these Proceedings). The details of the early SFHs of tiny dwarf galaxies can shed light, in particular, on the role in galaxy formation of the reionization which occurred at high redshift. Isolated dwarfs are ideal probes since their evolution is not complicated by environmental effects owing to the vicinity of the Milky Way or M31. Studying galaxy formation and evolution from Local Group galaxies. 251 Fig. 2. CMDs obtained with the ACS on board HST for four isolated Local Group dwarf galaxies. In order to give a first indication of the range of ages and metallicities present in each galaxy, isochrones from [16] (scaled solar, overshooting set), with the ages and metallicities indicated in the labels, have been superimposed. The locus of the zero-age horizontal-branch has also been represented. Distance moduli of (m − M )0 = 24.4, 24.0, 24.45 and 24.7 and reddenings of E(B-V)=0.04, 0.05, 0.03 and 0.03 for IC1613, LGS3, Cetus and Tucana, respectively, have been adopted to transform the isochrones to the observational plane. Determinations of the SFH of each system are under way through comparison with synthetic CMDs. 2.4 The study of stellar population gradients In all these programmes, in addition to the derivation of accurate and detailed SFHs, we pay special attention to the study of the stellar population gradients. Their presence in dwarf galaxies, with the youngest population concentrated towards their centre, is well known (e.g. [1], [4], [13]). In the Milky Way satellites, for which we have CMDs reaching the oldest main-sequence turnoffs, the nature of these gradients can be investigated in detail; here the difficulty is due to the large areas that need to be surveyed. For example, in the case of the 252 C. Gallart Magellanic Clouds, their total extension and the presence or not of an old halo is still a matter of debate. In the case of the more distant dIrr galaxies, the actual nature of the outer older structure, and its extension, is still uncertain due to the faintness of the stars that need to be measured. Recent studies show the existence of both young and old populations in the central parts of dIrr galaxies, while young stars gradually disappear towards the outer regions. But they don’t offer enough proof of the actual nature of these extended structures, and in particular, whether they represent a true halo population (i.e., old and tracing the initial conditions of galaxy formation). Recent results for the LMC ([10]) and Phoenix ([12]), using CMDs that reach the oldest main-sequence turnoffs indicate that these extended regions are not exclusively old, and that a smooth age gradient exists from the centre of the galaxy to its outer parts. This suggests an outside–in formation scenario, contrary to what seems to happen to the discs of spiral galaxies. With the CMDs of isolated dwarfs shown in the previous Section, it will be possible to investigate in detail the age distribution of the stellar populations in each galaxy as a function of radius, thus sheding new light on the possible formation mechanisms. An additional key diagnostic is the kinematics of stars (e.g., [18]) of different ages, which will provide information of the dynamical evolution of the galaxy, possibly indicating the presence or otherwise of differentiated disc–halo structures, or the presence of distinct kinematic entities, possibly originating in the accretion of smaller systems (according to the predictions of hierarchical galaxy formation models). This is a field still to be explored for dwarf galaxies outside of the Milky Way satellite system, and Flames at the VLT and OSIRIS at the GTC will be key instruments for this purpose. 3 Metallicities using the Ca II triplet. We have obtained (PhD, R. Carrera) a new calibration of the CaII triplet strength in red giant branch (RGB) stars as a function of metallicity, which is valid for a higher range of ages (13 ≤ Age(Gyr) ≤ 0.25) and metallicities (−2.2 ≤ [Fe/H] ≤ +0.5) than previously published calibrations (see [7] for the most recent one). This calibration has been used to obtain metallicities for a large number of stars in different fields of the LMC and the SMC (see the contribution by R. Carrera et al. in these Proceedings for details). With the GTC and OSIRIS we plan to extend this type of work to the remaining galaxies in the Local Group, situated at a distance of ' 1 Mpc. Candidate RGB stars are in the magnitude range I '21–21.5. They will densely populate the OSIRIS field of view ('30 stars per sq. arcminute) to allow us efficient use of multiobject spectroscopy. The high brightness of the sky in the CaII triplet region implies that the possibility of micro-slit nod-and-shuffle will be important for this project. Studying galaxy formation and evolution from Local Group galaxies. 253 4 Variable stars Variable stars can complement the information offered by CMDs to interpret the stellar populations of a galaxy. In particular, RR Lyrae reveal the presence of a very old ('10 Gyr) stellar population, while short-period classical Cepheids and anomalous Cepheids are tracers of populations up to a few hundred Myr old and a few Gyr old respectively ([8]). Using variable stars as stellar population indicators is especially important when it is not possible to obtain CMDs reaching the oldest main-sequence turnoffs. The ACS data mentioned in Section 2 is also excellent for obtaining a census of the variable star population in each galaxy of the sample. Such a study is already under way (see the contribution by E. Bernard et al. in these Proceedings). The good sensitivity and relatively large field of OSIRIS at the GTC will allow us to carry on a systematic characterization of the variable star populations in Local Group galaxies (both isolated dwarfs and M31 dSph companions, and some strategic fields in the large M31 and M33 spirals). Such surveys will complement in a key way our ACS imaging project: with ACS, only a handful of galaxies will be studied, and in some cases only part of their total extent will be covered. To find RR Lyrae stars, the most challenging and interesting part of this project, we need to reach g0 ' 25.5 in relatively short exposure times. OSIRIS will allow us to do that in 10–15 minutes. In addition, its field of view is very well suited to cover most Local Group dwarf galaxies in one to a few fields. Acknowledgements: Support for this project is provided by the IAC (Project 3I1902), the Spanish Ministry of Science and Technology (AYA2004-06343), and the European Structural Funds. The Stellar populations in Galaxies group at the IAC is currently composed of A. Aparicio, E.J. Bernard, R. Carrera, I. Drozdovsky, A. Marı́n-Franch, I.P. Meschin, M. Monelli, N.E.D. Noël, A. Rosenberg and myself. I thank the co-investigators of the projects discussed in this paper for allowing me to show results in advance of publication. In particular (and in addition to the IAC Group members quoted above) F. Pont, E. Hardy, P. Stetson and R. Zinn (LMC project), E. Costa and R. Méndez (SMC project), and the LCID Team (Local Constraints from Isolated Dwarfs Team: A. Aparicio, E.J. Bernard, G. Bertelli, S. Cassisi, A.A. Cole, P. Demarque, A. Dolphin, I. Drozdovsky, H.C. Ferguson, L. Mayer, M.L. Mateo, M. Monelli, J. Navarro, S.L. Hidalgo, F.J. Pont, E.D. Skillman, P.B. Stetson & E. Tolstoy). References 1. A. Aparicio, J. (1997) 2. A. Aparicio, R. 3. A. Aparicio, C. 4. A. Aparicio, N. J. Dalcanton, C. Gallart, D. Martinez-Delgado: AJ, 114, 1447 Carrera, D. Martı́nez-Delgado: AJ, 122, 2524 (2001) Gallart: AJ, 128, 1465 (2004) Tikhonov, I. Karachentsev: AJ, 119, 177 (2000) 254 C. Gallart 5. R. Carrera, A. Aparicio, D. Martı́nez-Delgado, J. Alonso-Garcı́a: AJ, 123, 3199 (2002) 6. M.-R. L. Cioni, H. J. Habing, F. P. Israel: A&A, 358, L9 (2000) 7. A. A. Cole, T. A. Smecker-Hane, E. Tolstoy, T. L. Bosler, J. S. Gallagher: MNRAS, 347, 367 (2004) 8. C. Gallart, A. Aparicio, W. L. Freedman, B. F. Madore, D. Martı́nez-Delgado, P. B. Stetson: AJ, 127, 1486 (2004) 9. C. Gallart, W. L. Freedman, A. Aparicio, G. Bertelli, C. Chiosi: AJ, 118, 2245 (1999) 10. C. Gallart, P.B. Stetson, E. Hardy, F. Pont, R. Zinn: ApJ, 614, L109 (2004) 11. C. Gallart, M. Zoccali, A. Aparicio: ARAA, 43, 387 (2005) 12. S. L. Hidalgo, A. Aparicio, D. Martı́nez-Delgado: ApJ, submitted (2006) 13. D. Harbeck et al: AJ, 122, 3092 (2001) 14. J. A. Holtzman et al: AJ, 118, 2262 (1999) 15. N. E. D. Noël, C. Gallart, E. Costa, R. Méndez: AJ, submitted (2006) 16. A. Pietrinferni, S. Cassisi, M. Salaris, F. Castelli: ApJ, 612, 168 (2004) 17. T. A. Smecker-Hane, A. A. Cole, J. S. Gallagher, P. B. Stetson: ApJ, 566, 239 (2002) 18. E. Tolstoy et al: ApJ, 617, L119 (2004) 19. D. Zaritsky, J. Harris, E. K. Grebel, I. Thompson: ApJ, 534, L53 (2000) Gaia: A major step in the knowledge of our Galaxy J. Torra on behalf of the Gaia Group Dpt. d’Astronomia i Meteorologia, IEEC-Univ. de Barcelona, Avda. Diagonal 647, E-08028 Barcelona, Spain, jordi@am.ub.es Summary. The ESA’s space mission Gaia will offer a new view of our Galaxy. Based on the success of Hipparcos, the huge amount of astrometric and astrophysical data it will provide along with its unprecedented quality, will allow new and well grounded studies on structure, kinematics and evolution of our Galaxy. We review in this presentation the main features and capabilities of the Gaia mission after the start of the industrial phase. The plans for future development and exploitation are also presented. 1 Introduction Gaia will create the largest and most precise three dimensional survey of our Galaxy and beyond by providing unprecedented positional, radial velocity and spectroscopic data for about one billion stars in our Galaxy and throughout the Local Group ([4], [17]). Gaia launch is scheduled for December 2011 and mission end five years later. The detailed design and construction phase (B2) started in February 2006. The knowledge of the structure and evolution of our Galaxy has ran in parallel to the availability of large surveys. As new data are published new constraints to the galactic models and new values (scale heights and lenghts, densities, etc.) for the characterization of galactic structures are determined. In the recent years, after the enormous success of the Hipparcos catalogue ([16]), mainly by its accurate parallaxes, that superseeded some of the at the epoch ”best” catalogues such as the Nearby Stars Catalogue ([11]), several new catalogues have appeared. Among them we can mention the Tycho II ([10]), a product of the Hipparcos mission too, that with its 2.5 milion stars (positions at 60 mas, proper motions at 2.5 mas/yr, BT and VT photometry at 0.10 mag) has been largely used and cross-matched with other observational catalogues (i.e. ROSAT) to get kinematical and astrophysical data for several purposes. Let us mention its success in studying the moving groups, although a number of citations, in different fields, can be found elsewhere in this colloquium. 256 J. Torra on behalf of the Gaia Group Among the largest catalogues containing astrophysical information we should mention UCAC2 (USNO CCD Astrograph catalogue) whose final version forecast for 2007 will contain positions up to 70 mas at R = 16 plus proper motions derived using Hipparcos, Tycho-II, and AC2000 and remeasures of AGK2 plates. It has been cross-matched with 2MASS thus providing photometry for most of its 48 million stars. Unfortunately large catalogues like USNO B1.0, offering position (200 mas) and magnitudes for some billion stars and designed for specific purposes like the guidance of HST or JWST, are far of being of general use. Since the publication of IRAS data, infrared surveys have had a large impact in the understanding of our Galaxy. 2MASS ([19]) providing J,H, K photometry for 300 milion objects, the 2 micron Galactic survey ([9]), or Denis ([7]) giving I,J,K, are good examples of catalogues allowing the deep studies in the directions of the galactic center. We must remark to end, the availability of non stellar surveys like the COBE-DIRBE photometry from wich accurate extinction maps have been derived. Gaia will increase the precision of the best astrometric catalogues by twothree orders of magnitude increasing by a factor 104 the number of objects. Furthermore Gaia will offer a unprecedented approach for the knowledge of our Galaxy: it will provide the astrophysical data necessary to classify all the objects and the radial velocities for a large fraction of them, thus mapping both the phase space and the distribution of physical properties. Let us remark that this determination of radial velocities and flux data is not only a goal of the mission but also a request. To reach the microarcsecond level of the astrometric precision correction of chromatic effects have to be considered and the astrometric model of the observations must include radial velocity data. 2 GAIA: The Scientific Case 2.1 Origen, Formation and Evolution of the Galaxy The main goal of the Gaia mission is to produce the set of homogeneous and accurate data needed to perform the most deep and well grounded study of the structure, origin, formation and evolution of our Galaxy. To reach this goal a large and unbiased sample of positional, kinematical and astrophysical data of a statistically significant sample of the stellar content of the Galaxy is needed. At the end of the mission Gaia will provide: • positions and proper motions with precisions better than 20 µas and 20 µas yr−1 at V = 15 • parallax data with 20 µas at V = 15, that is, 20% precision in distance at 10 kpc • stellar atmospheric parameters (temperature, gravity, chemical composition) for all the stars up to V = 18 Gaia: A major step in the knowledge of our Galaxy 257 • radial velocities of about 15 kms−1 precission at V = 17 These figures have been settled by ensuring the equilibrium among the characteristics of the main tracers of galactic populations (i.e. horizontal branch stars for the bulge, K-giants for the thick disk, M-giants for the warp) and the technical constraints of the mission. The list of scientific goals of Gaia has been discussed extensively in [4] and [5]. Let us briefly comment as examples three topics on which Gaia data will be fundamental: • The star formation To determine the star formation history of the Galaxy we need to know the evolution of the star formation rate, as well as the total number of stars formed eveywhere in the Galaxy. This information along with the kinematic data -giving clues for the merging of stellar systems-, and the chemical abundance data -related to the accretion processes-, will allow the determination of the evolution of the Milky Way. The analysis of Gaia results with powerful statistical methods will provide for the first time a quantitative determination of the formation history of our Galaxy. • The stellar astrophysics The determination of the luminosity of a star is based on the knowledge of its distance and of the interstellar extinction. Distances can be estimated through trigonometric parallaxes and extinction through photometry. Distances accurate better than 10% can extend up to 10 kpc, thus including the galactic center, spiral arms, halo etc. On the other hand, the Gaia’s limiting magnitude will allow the determination of acurate distance for white dwarfs and brown dwarfs. This means that luminosities will be measured for all the stellar types along the HR diagram thus establishing hard constraints for the stellar structure and evolutionary models. • The open clusters The number of open cluster in the Lynga catalogue ([15]) was of 1200, of which some 400 had accurate although heterogeneous data derived from photometry and very few had space velocities. After Hipparcos and Tycho catalogues new detections have been made. [6] published a new list with new 356 entries. In 2006, [18] have published new data for some of these clusters and have added 130 open clusters to the lists. On the other hand, [3] had detected some 500 embedded infrared clusters in the 2MASS catalogue. Gaia will be abble to identify members of almost all clusters closer than 5 kpc. This young and intermediate-age disc tracers by its uniformity in composition and age are excellent empirical references for the study of stellar evolution, star formation and IMF determination. On the other hand they allow the kinematic study of structures like the Gould’s belt or the spiral arms. 258 J. Torra on behalf of the Gaia Group 3 Instrument Description The Gaia satellite and mission concept where described in [4]. In order to fit in the Soyuz-Fregat launcher and to satisfy budgetary conditions it was modified and resized in 2004, and once the project has entered the industrial phase a new design has been done although the basics of the satellite remain unchanged: Gaia is a revolving scanning satellite that superposes in a unique focal plane the FoVs of two telescopes (apertures are 1.45 × 0.45 m2 ) pointing to two directions separated by the basic angle of 106 degrees. The images obtained cross a large array of CCDs operated in TDI mode placed in the common focal plane. Fig. 1. The configuration of the Gaia instrument The main and important differences with the previous designs are: • the substitution of the so called SPECTRO instrument, where photometry was obtained using a number of filters and radial velocities determined through a slitless grating spectrograph, mounted on the focal plane of a dedicated telescope, by two low dispersions prisms(a red one with 7-15 nm/pixel dispersion and blue one with 4 - 34 nm/pixel) and a grating spectrographs operating in the CaII triplet with R = 11500. All these three spectrographs are fed by the astrometric telescopes as can be seen in Fig. 1 • the redimensioning of the focal plane which now is constituded (see Fig. 2 by 106 CCD (pixel size 10×30 µm2 , 4500 pixels along-scan and 1966 pixels across-scan per CCD), 76 of them are devoted to the detection and astrometric measurements, 2 columns of 7 rows each are dedicated to the red and blue spectrographs while three columns of four rows are for radial Gaia: A major step in the knowledge of our Galaxy 259 velocity measurements. The remaining four CCD are for wavefront sensors and basic angle monitoring. Fig. 2. The layout of the focal plane 3.1 Astrometry Gaia observation are not biased by selection effects. The objects to be observed are detected on-board by the two star-mappers and once a detection is confirmed, a window is defined and the object tracked along the focal plane. These window data are treated on-ground to determine the centroid of the observed samples, to cross-match the different observation (84 on average of a given star) and to enter the Global Iterative Solution (GIS) ([14]) where the astrometric as well as the calibration and attitude data are determined at the same time. The end-of-mission astrometric accuracies depends on Gaia’s scanning law (SL) properties, thus implying dependence respect to the direction on the sky. The properties of Gaia’s optical and detector system are such that for stars brighter than ∼12 mag photon noise is negligible. The end-of-mission astrometric accuracies for these stars will amount to a few µas, and the GIS is a smaller contribution. For magnitudes between 12 and 20, the expected accuracies range is from 20 - 25 µas at 15-th magnitude to a few hundred µas at 20-th magnitude. At a given magnitude, astrometric accuracy also depends on apparent star colour. Generally, redder stars have smaller astrometric errors. 260 J. Torra on behalf of the Gaia Group 3.2 Spectrophotometers As said above photometry has to be derived from the red and blue spectrographs. Deep analysis was done, using previous Gaia designs, to optimize pass-bands [12]. The problem now is quite more complicated. The convolution of the square aperture with the dispersion of the BP/RP system gives images on which synthetic aperture photometry and calibration has to be performed. Other effects like image blurring by TDI or attitude irregularities can do it even more difficult [13]. Nonetheless as Gaia observes each star many times and the number of observation in a given CCD is very large, that will result in a complicated model of the observation. The same philosophy applies to merging of spectra by the fact that the different observations of a given object are performed scanning the sky in different directions. 3.3 Radial Velocity Spectrometer The Radial Velocity Spectrometer (RVS) will operate at 847-874 nm, with a resolution of about 11500. An average of ∼40 spectra per star over the 5 years mission will be collected from which a final accuracy ∼ 1 km/s at the bright end and ∼10-15 km/s at magnitude G = 17, the faint end is expected. Thus, about 100-150 million stars will have complete data in the phase space (position and velocities). In addition to the information on star’s motion, RVS will bring an excellent physical characterization of bright stars. Examples are: • Multi-epoch radial-velocity information will be used to characterize double and multiple systems. Gaia will provide masses and radii accurate to a few per cent for thousands of eclipsing binaries. • The RVS will monitor the radial motions of the outer layers of pulsating stars. It will provide pulsation curves for RR Lyrae, Cepheids and Mira variables up to the 14th magnitude. • Individual abundances of key chemical elements, e.g. Ca, Mg and Si for all stars up to 12th magnitude are expected. Furthermore, the Diffuse Interstellar Band (DIB) placed at 862 nm is well inside the wavelength range of RVS thus providing a significant contribution to the derivation of the 3D interstellar reddening map. 4 2006-2016: complementary data and pre-launch modelling Besides the technical exigences and the challenge posed by the treatment of the Gaia data other issues that must be considered prior to the launch and up to the end of the mission are: Gaia: A major step in the knowledge of our Galaxy 261 • Techniques for the construction of dynamical Galaxy Models, considered essential infrastructure that should be put in place before Gaia flies, are being developed by the european scientific community ([5]). • The level of precision requested makes necessary an effort in the modelling of the observations. A relativistic model of the astrometric observations has been set up and several effects like light-travel time have been considered. On the other hand the use of Gaia to determine some relativistic parameters is under study and must be defined before launch. [1] and [2]. • From the point of view of spectrophotometry a detailed modelling of critical effects like charge transfer inefficiencies and calibratioon of the BP/RP response [13] must be considered. • Low resolution spectra on the focal plane will be highly overlapped not only in crowded regions like the Baade’s window, where the sky density rise above 3106 stars, but also in all the areas surrounding and inside the galactic bulge. The advantadge of having different scan directions at each passage shall be explored. Nonetheless in some cases overlapping is unavoidable and Gaia will need data coming from on ground observations with large telescopes. 5 Data Processing and Analysis It has been always recognized that the treatment of the Gaia data has a very demanding process. A first prototype was created, starting in 2000 [20], capable to perform the most demanding operation for the astrometric point of view. The GDAAS prototype demonstrated [8] for the first time the GIS approach and gave quantitative and qualitative estimates for the development of an operational system. Processes similar to GIS are considered now for photometry and spectroscopy. We have to bear in mind that the Gaia Data treatment will involve some 1021 flops against a database ∼ 1-2 PB. To cope with this challenging problem ESA has recently issued and AO to the scientific community to build up a consortium to take care of all the aspects of the data treatment and management. The Gaia Data Processing and Analysis Consortium (DPAC) created on June ([21]) from the former Gaia Working Groups, is, to our knowledege, the only candidate to do that task. DPAC is organized in nine Coordination Units each one taking care of a particular aspect of the data treatment [21] Acknowledgements: Project funded by Spanish MCyT: PNE2003-04352, ESP200524356-E, PNE2006-13855-C02-01 and ESP2006-26356-E. References 1. G. Anglada-Escudé, J. Torra: A&A 449, 1281 (2005) 262 J. Torra on behalf of the Gaia Group 2. G. Anglada-Escudé, J. Torra, S.A. Klioner et al: this volume (2007) 3. E. Bica,, C.M. Dutra, B. Barbuy: A&A 397, 177 (2003) 4. K.S. de Boer, G. Gilmore, E. Hoeg et al: GAIA, Composition, Formation and Evolution of the Galaxy, ESA-SCI(2000)4, Noordwijk, Netherlands: ESA Publications Division (scienfitic case on-line at http://astro.estec.esa.nl/GAIA) (2000) 5. J. Binney: Modeling the Galaxy for Gaia. In: The Three-Dimensional Universe with Gaia, ESA SP-576, ed by. C. Turon, K.S. O’Flaherty, M.A.C. Perryman pp 89 (2005) 6. W.S. Dias, B.S. Alessi, A. Moitinho et al: A&A 389, 871 (2002) 7. N. Epchtein, E. Deul, S. Derriere, et al: A&A 349, 236 (1999) 8. F. Figueras, B. López Martı́, C. Fabricius, C.et al: Data Processing: Testing of Core Tasks. In: The Three-Dimensional Universe with Gaia, ESA SP-576, ed by. C. Turon, K.S. O’Flaherty, M.A.C. Perryman pp 369 (2005) 9. F. Garzón, P.L. Hammersley, T. Mahoney et al: MNRAS 264, 773 (1993) 10. E. Hog, C. Fabricius, V.V. Makarov et al: A&A 357, 367 (2000) 11. H. Jahreiss, R. Wielen: The impact of Hipparcos on the Catalogue of Nearby stars. The stellar luminosity function and local kinematics. In: HIPPARCOS ’97, ESA SP-402, ed by B. Battrick, M.A.C. Perryman and P.L. Bernacca, pp 675 12. C. Jordi, E. Hoeg, A.G.A. Brown et al: MNRAS 367, 290 (2006) 13. C. Jordi, J.M. Carrasco, E. de Amores et al: this volume (2007) 14. Lindegren, L.: The Astrometric Instrument of Gaia: Principles. In: The ThreeDimensional Universe with Gaia, ESA SP-576, ed by. C. Turon, K.S. O’Flaherty, M.A.C. Perryman pp 29 (2005) 15. G. Lynga: Catalogue of open cluster data. Computer based catalogue available through CDS, Strasbourg, France, 5th edition (1987) 16. Perryman, M.A.C., ESA: The Hipparcos and Tycho catalogues. Astrometric and photometric star catalogues derived from the ESA Hipparcos Space Astrometry Mission, ESA SP-1200, Noordwijk, Netherlands: ESA Publications Division (1997) 17. M.A.C. Perryman, K.S. de Boer, G. Gilmore, G.et al: A&A 369 (2001) 18. A.E. Piskunov, N.V. Kharchenko, S. Rser,et al: A&A 445, 545 (2006) 19. M.F. Skrutskie, S.E. Schneider, R. Stiening, et al: The Two Micron All Sky Survey (2MASS): Overview and Status. In: The Impact of Large Scale Near-IR Sky Surveys, ed by F. Garzon et al. (Dordrecht: Kluwer Academic Publishing Company) pp 25 (1997) 20. J. Torra, J., F. Figueras, C. Jordi, et al: Design and Overview of the Data Processing. In: The Three-Dimensional Universe with Gaia, ESA SP-576, ed by. C. Turon, K.S. O’Flaherty, M.A.C. Perryman pp 361 (2005) 21. J. Torra, F. Figueras, C. Jordi, C. et al: this volume (2007) Cepheus A, a laboratory for testing and opening new theories on high-mass star formation J.M. Torrelles1, N.A. Patel2 , S. Curiel3 , G. Anglada4 , and J.F. Gómez4 1 2 3 4 Instituto de Ciencias del Espacio (CSIC) and Institut d’Estudis Espacials de Catalunya, Facultat de Fı́sica, Planta 7a, Universitat de Barcelona, Av. Diagonal 647, E-08028 Barcelona, Spain, torrelles@ieec.fcr.es Harvard-Smithsonian Center for Astrophysics, 60 Garden St., Cambridge, MA 02138, USA npatel@cfa.harvard.edu Instituto de Astronomı́a, Universidad Nacional Autónoma de México, Apartado Postal 70-264, D.F. 04510, México scuriel@astroscu.unam.mx Instituto de Astrofı́sica de Andalucı́a (CSIC), Apartado 3004, E-18080 Granada, Spain guillem@iaa.es, jfg@iaa.es Summary. Cepheus A is one of the closest (725 pc distance) high-mass star-forming regions. At the center of a high-density molecular core seen in ammonia lines, within a ' 20” (' 15000 AU) radius, there is a cluster of sixteen compact radio continuum sources observed at centimeter wavelengths, some of them associated with phenomena such as Herbig-Haro (HH) objects, molecular outflows, jets, disks, masers, and strong magnetic fields, all of which are signatures of the first steps of the evolution of very young stars. In this contribution, we will review some of the past and most recent observations of Cepheus A, a laboratory for testing and opening new puzzling questions related to the early stellar evolution. 1 Introduction Young stellar objects (YSOs) are characterized by powerful and highly collimated jets, Herbig-Haro (HH) emission, molecular outflows, maser emission (e.g., H2 O, OH, CH3 OH), radio continuum emission, and strong magnetic fields (e.g., [1], [2], [25], [33]). All these phenomena can be explained within a general scenario of an accretion disk surrounding a protostar, ejecting collimated outflows through the poles of the disk and its subsequent interaction with the ambient medium. This scenario seems to be valid, as a first approach, from low to intermediate and high-mass stars (e.g., [24], [11], [18]). However, our knowledge of the earliest stages of massive stars (≥ 8 M ) has been limited by the lack of data to study these objects with sufficient angular resolution (sub-arcsecond resolution) and sensitivity, given that they are located at distances typically higher than 1 kpc. 264 Torrelles, Patel, Curiel, Anglada, & Gómez Although in the last few years there has been a growing evidence that high-mass stars may form through accretion of material from a rotating circumstellar disk (in a similar way as low-mass stars form; see [11], [7], [23], [34], [20], [5], [3], [4], [19], [9], [32]), rather than through the merging of several low-mass stars (e.g., [6]), there is still a deficit in the detection of jets and disks in high-mass YSOs at scales of ≤ 1000 AU where relevant physical (e.g., accreting processes, rotating motions, outflow collimation) are expected to occur. In this contribution, we will review some of the past and most recent observations of Cepheus A. In particular, we will review the results obtained by [19] and [9] which show the presence of a disk-jet system associated with a high-mass YSO, supporting that massive stars form as low-mass stars do. In addition, we will also show the case of the expanding spherical water maser bubble ejected from a young star ([29, 30, 33]), a phenomenon that is not predicted by current theories on star formation. 2 A disk of dust and molecular gas around a high-mass (∼ 15 M) young stellar object Cepheus A is one of the closest high-mass star forming region (725 pc; [16, 22]) exhibiting many of the phenomena related with YSOs (see [13], and references therein). At the center of a high-density molecular core there are multiple radio continuum sources and strong H2 O and OH masers ([15], [21], [14], [12], [27], [8], [17], [33]). Several of the HW (Hughes & Wouterloot) radio continuum sources detected in this region are excited by internal sources, while others appear to be shock-excited, delineating the edges of high-density molecular cores seen in ammonia line emission ([12], [28], [31], [8]). HW2, which is the brightest radio continuum source of the region, is a thermal biconical radio jet excited by a massive star of ∼ 15 M ([21]). This jet is powering the more extended bipolar molecular outflow seen in HCO+ ([13]). Through Very Large Array (VLA) multiepoch observations, [9] have measured large proper motions in the components of the radio jet, with the two main components of the jet moving away from the central source in nearly opposite directions with velocities of ∼ 500 km s−1 (Figure 1). In addition, subarcsecond Submillimeter Array (SMA) observations carried out by [19] have revealed a flattened disk-like structure in both dust and CH3 CN line emission of ∼ 600 AU in size and mass ∼ 1-8 M oriented perpendicular to, and spatially peaking at the center of the HW2 jet (Figure 2), just as is the case with low-mass stars. All these observations strongly suggest a disk interpretation for the flattened structure seen in dust and CH3 CN, giving support to theoretical models of high-mass star formation via an accretion process occurring in a disk around the protostar (e.g., [10]), as low-mass stars do (e.g., [24]). Cepheus A, a laboratory for testing and opening new theories 265 Fig. 1. VLA contour maps of the HW2 thermal biconical jet at 3.6 cm. The triangle and star indicate the positions of the northeast and southwest knots for the 1999 epoch, respectively. The proper motions of these knots are evident in subsequent epochs. The half-power contour of the beam is shown in each bottom left-hand corner (Figure from [9]). 3 Spherical episodic ejection of material from a young star VLA observations of water maser emission made with an angular resolution of 0.“08 show four cluster of masers spread over ∼ 500 around the HW2 radio jet ([26, 27]). One of these is associated with the jet, two are associated with the radio continuum sources HW3b and HW3d, and the third one is located ' 0.00 7 (500 AU) south of HW2. Very Long Baseline Array (VLBA) three-epoch water maser observations with ' 0.5 mas angular resolution have shown that this last cluster forms in the sky a bright arc-like structure of ' 100 mas size (' 72 AU), persisting in the three epochs. The arc of masers is extremely well fitted 266 Torrelles, Patel, Curiel, Anglada, & Gómez Fig. 2. Dust continuum and CH3 CN line emission at ∼ 330 GHz from the disk associated with the high-mass star object HW2 as observed with the Submillimeter Array (beam size ' 0.00 75). The elongation in both dust (gray scale) and CH3 CN emission (thick contours) is nearly perpendicular to the biconical thermal radio jet (oriented and with proper motions in the northeast-southwest direction; thin [3.5 cm] and white [1.3 cm] contours, see also Fig. 1), supporting the disk-jet interpretation for HW2. The deconvolved disk radius is ∼ 330 AU with a mass of ∼ 1-8 M . The SMA beam is shown in the lower left corner (Figure from [19]). Cepheus A, a laboratory for testing and opening new theories 267 by a circle of 62 AU radius to an accuracy of one part in a thousand. Moreover, this arc structure is constituted by a string of smaller linear structures of maser spots with sizes ' 0.4-1 AU and different orientations, all of which are tangential to the arc curvature ([29, 30]). Proper motions of the water masers indicate a uniform expansion of 9 km s −1 perpendicular to the arc. The high degree of symmetry of the arc as well as its dynamical age of 33 yr suggest that it represents the limb brightened parts of a singular, short lived, and episodic spherical ejection (explosive event) driven by a YSO located at the center of the circle which fits the arc structure, with the smaller linear structures being most likely highly flattened surfaces defining shock fronts. This YSO was detected in subsequent studies by [8] at 3.6 cm continuum, although its nature is still unknown. In addition, [33], through VLBA linear and circular polarization observations, have measured the strength of the magnetic field in this arc structure (∼ 30-130 mG), with a direction along the shell expansion direction, radial from the central embedded YSO. The physical relevance of all these VLBA results is found in the fact that isotropic ejections are difficult to explain within the current paradigm of star formation, where we expect to observe collimated outflows due to the presence of circumstellar disks. At the moment, the origin of this spherical expanding bubble and the role of the magnetic field in that kind of ejection remains unknown, but it could represent new constraints for theories on how stars evolve in their early stages. References 1. Anglada, G.: RevMexAASC, 1, 67 (1995) 2. Anglada, G.: in Radio emission from the stars and the sun, edited by A. R. Taylor and J. M. Paredes, ASP Conference Series, 93, 3 (1996) 3. Beltrán, M. T., Cesaroni, R., Neri, R., Codella, C., Furuya, R. S., Testi, L., Olmi, L.: ApJ, 601, L187 (2004) 4. Beltrán, M. T., Cesaroni, R., Codella, C., Testi, L., Furuya, R. S., Olmi, L.: Nature, 443, 427, (2006) 5. Beuther, H., Schilke, P., Sridharan, T. K., Menten, K. M., Walmsley, C. M., Wyrowski, F.: A&A, 383, 892 (2002) 6. Bonnell, I. A., Bate, M. R.: MNRAS, 336, 659 (2002) 7. Cesaroni, R., Felli, M., Jenness, T., Neri, R., Olmi, L., Robberto, M., Testi, L., Walmsley, C. M.: A&A, 345, 949 (1999) 8. Curiel, S., et al.: ApJ, 564, L35 (2002) 9. Curiel, S., Ho, P. T. P., Patel, N. A., Torrelles, J. M., Rodrı́guez, L. F., Trinidad, M. A., Cantó, J., Hernández, L., Gómez, J. F., Garay, G., Anglada, G.: ApJ, 638, 878 (2006) 10. De Buizer, J. M., Osorio, M., Calvet, N.: ApJ, 635, 452 (2005) 11. Garay, G., Lizano, S.: PASP, 111, 1049 (1999) 12. Garay, G., Ramı́rez, S., Rodrı́guez, L. F., Curiel, S., Torrelles, J. M.: ApJ, 459, 193 (1996) 268 Torrelles, Patel, Curiel, Anglada, & Gómez 13. Gómez, J. F., Sargent, A. I., Torrelles, J. M., Ho, P. T. P., Rodrı́guez, L. F., Cantó, J., Garay, G.: ApJ, 514, 287 (1999) 14. Hughes, V. A., Cohen, R. J., Garrington, S.: MNRAS, 272, 469 (1995) 15. Hughes, V. A., Wouterloot, J. G. A.: ApJ, 276, 204 (1984) 16. Johnson, H. L.: ApJ, 126, 121 (1957) 17. Martı́n-Pintado, J., Jiménez-Serra, I., Rodrı́guez-Franco, A., Martı́n, S., Thum, C.: ApJ, 628, L61 (2005) 18. Osorio, M., Lizano, S., D’Alessio, P.: ApJ, 525, 808 (1999) 19. Patel, N. A., Curiel, S., Sridharan, T. K., Zhang, Q., Hunter, T. R., Ho, P. T. P., Torrelles, J. M., Moran, J. M., Gómez, J. F. G., Anglada, G.: Nature, 437, 109 (2005) 20. Ridge, N. A.,Moore, T. J. T.: A&A, 378, 495 (2001) 21. Rodrı́guez, L. F., Garay, G., Curiel, S., Ramı́rez, S., Torrelles, J. M., Gómez, Y., Velázquez, A.: ApJ, 430, L65 (1994) 22. Sargent, A. I.: ApJ, 218, 736 (1977) 23. Shepherd, D., Claussen, M. J., Kurtz, S. E.: Science, 292, 1513 (2001) 24. Shu, F. H., Adams, F. and Lizano, S.: ARA&A, 25, 23 (1987) 25. Torrelles, J. M., Gómez, J. F., Anglada, G., Estalella, R., Mauersberger, R., Eiroa, C.: ApJ, 392, 616 (1992) 26. Torrelles, J. M., Gómez, J. F., Garay, G., Rodrı́guez, L. F., Curiel, S., Cohen, R., Ho, P. T. P.: ApJ, 509, 262 (1998) 27. Torrelles, J. M., Gómez, J. F., Rodrı́guez, L. F., Curiel, S., Ho, P. T. P., Garay, G.: ApJ, 457, L107 (1996) 28. Torrelles, J. M., Ho, P. T. P., Rodrı́guez, L. F., Cantó, J.: ApJ, 305, 721 (1986) 29. Torrelles, J. M., Patel, N. A., Gómez, J. F., Ho, P. T. P., Rodrı́guez, L. F., Anglada, G., Garay, G., Greenhill, L., Curiel, S., Cantó, J.: Nature, 411, 277 (2001) 30. Torrelles, J. M., Patel, N. A., Gómez, J. F., Ho, P. T. P., Rodrı́guez, L. F., Anglada, G., Garay, G., Greenhill, L., Curiel, S., Cantó, J.: ApJ, 560, 853 (2001) 31. Torrelles, J. M., Verdes-Montenegro, L., Ho, P. T. P., Rodrı́guez, L. F., Cantó, J.: ApJ, 410, 202 (1993) 32. Trinidad, M. A., Curiel, S., Torrelles, J. M., Rodrı́guez, L. F., Migenes, V., Patel, N. A.: AJ, 132, 1918 (2006) 33. Vlemmings, W. H. T., Diamond, P. J., van Langevelde, H. J., Torrelles, J. M.: A&A, 448, 597 (2006) 34. Zhang, Q., Hunter, T. R., Brand, J., Sridharan, T. K., Molinari, S., Kramer, M. A., Cesaroni, R.: ApJ, 552, L167 (2001) Session VI Sun and planetary systems A Look into the Guts of Sunspots L.R. Bellot Rubio Instituto de Astrofı́sica de Andalucı́a (CSIC), Apdo. 3004, 18080 Granada, Spain, lbellot@iaa.es Summary. Advances in instrumentation have made it possible to study sunspots with unprecedented detail. New capabilities include imaging observations at a resolution of 0.001 (70 km on the sun), spectroscopy at ∼ 0.002, and simultaneous spectropolarimetry in visible and infrared lines at resolutions well below 100 . In spite of these advances, we still have not identified the building blocks of the penumbra and the mechanism responsible for the Evershed flow. Three different models have been proposed to explain the corpus of observations gathered over the years. The strengths and limitations of these models are reviewed in this contribution. 1 Introduction Sunspots were the first celestial objects known to harbor magnetic fields, a discovery made by Hale in 1908 [15]. One year later, Evershed described a nearly horizontal plasma outflow in sunspot penumbrae [14]. This flow produces the so-called Evershed effect: redshifted spectral lines in the limb-side penumbra and blueshifts in the center-side penumbra (Fig. 1). As seen in continuum images, the penumbra is formed by bright and dark filaments oriented radially. Observations have revealed a close relationship between the filamentary structure of the penumbra, its magnetic field, and the Evershed flow. The penumbra exhibits a complex magnetic topology, with fields of different strengths and inclinations interlaced both vertically and horizontally (see [37] and [3] for reviews). The more inclined fields channel the Evershed flow, while the more vertical fields are not associated with significant mass motions. In the inner penumbra, the magnetic field and the flow are directed upward [26, 34, 5, 7, 27], but in the outer penumbra one observes downward flows [26, 34, 32] along magnetic field lines returning back to the solar surface [44, 21, 5, 9, 17, 25]. The vertical interlacing of different magnetic field components with different velocities is responsible for the non-zero net circular polarization (NCP) of spectral lines emerging from the penumbra. These ingredients led to the concept of uncombed penumbra [39] (see also [43] and [19]). Basically, an uncombed penumbra consists of nearly hori- 272 L.R. Bellot Rubio continuum intensity 0.4 0.6 0.8 mean line-wing velocity [km/s] 1.0 -2 -1 0 1 2 Fig. 1. AR 10905 as observed with IBIS at the DST of NSO/Sac Peak Observatory on Aug 24, 2006. The spot was located 42o off the disk center. The spatial resolution is about 0.003. The observations were taken in the Fe i 709.0 nm line. Left: Continuum image. Right: Dopplergram derived from line-wing intensities. Positive velocities indicate blueshifts. The arrow points to disk center. Blueshifts in the center-side and redshifts in the limb-side penumbra are the signatures of the Evershed flow. Observations and data reduction courtesy of A. Tritschler and H. Uitenbroek. zontal magnetic flux tubes embedded in a stronger and more vertical ambient field. The tubes carry the Evershed flow, with the ambient field being essentially at rest. The uncombed penumbral model is supported by numerical simulations of interchange convection ([31] and references therein), but the detection of individual flux tubes in spectropolarimetric observations has proven elusive due to their small sizes (100–200 km in diameter). Recently, high-resolution (0.001–0.002) images taken with the Swedish 1-m Solar Telescope and the Dutch Open Telescope on La Palma have demonstrated that many penumbral filaments possess internal structure in the form of a dark core [29, 42]. The dark core is surrounded by two narrow lateral brightenings (Fig. 2, left), both of which are observed to move with the same speed and direction as a single entity. The fact that the various parts of dark-cored filaments show a coherent behavior have raised strong expectations that they could be the fundamental constituents of the penumbra, i.e., the flux tubes postulated by the uncombed model. Spectroscopy at 0.002 resolution suggests that the Evershed flow is stronger in the dark cores (Fig. 2, right) and that dark-cored filaments possess weaker fields than their surroundings close to the umbra [6]. Other than that, the magnetic and kinematic properties of dark-cored penumbral filaments remain unknown, so for the moment it is not possible to confirm or reject the idea that they represent individual tubes. In the meantime, alternative models of the penumbra have emerged: scenarios based on MIcro-Structured Magnetic Atmospheres (MISMAs; [23, 24]) A Look into the Guts of Sunspots 273 Fig. 2. Multi-line spectroscopy of dark-cored penumbral filaments at 0.002 resolution. The data were taken at the SST on April 29, 2005, and correspond to the centerside penumbra of AR 10756. Left: Slit-jaw image. The slit crosses four dark-cored filaments. Right: Intensity profiles of Fe i 557.6, Fe ii 614.9, and Fe i 709.0 nm along the slit. The dark cores (“DC”) are marked with small horizontal lines. Their large blueshifts are produced by Evershed flows directed upward. See [6] for details. and field-free gaps (the gappy penumbral model; [41]). These models try to explain the morphological and spectropolarimetric properties of the penumbra. They also claim to solve important problems of the uncombed model. In the following, the strengths and limitations of the different models are examined. 2 Competing penumbral models 2.1 Uncombed model As mentioned before, the uncombed model envisages the penumbra as a collection of small magnetic flux tubes embedded in an ambient field. The thermal, magnetic, and kinematic properties of the flux tubes and the ambient field (Fig. 3) have been determined from Stokes inversions that use two different magnetic atmospheres. These inversions [2, 5, 11, 9, 1, 10] have demonstrated that the uncombed model is able to explain the shapes of the polarization profiles of visible and infrared lines emerging from the penumbra at resolutions of ∼100 (see Fig. 4 for examples). Perhaps the most important achievement of the model, however, is that it quantitatively reproduces the NCP of visible [19, 10] and infrared [9, 22] lines, which are due to strong gradients or discontinuities of the atmospheric parameters (including velocities) along the line of sight. This success is not trivial, since the spatial distribution of the NCP is determined primarily by discontinuities of field inclination in the case of visible lines and discontinuities of field azimuth in the case of infrared lines [16, 33]. 274 L.R. Bellot Rubio Fig. 3. Radial variation of the field inclination (left) and field strength (center) in the penumbra of AR 8704 as derived from a two-component inversion of the Fe i lines at 1565 nm. Solid and dashed lines represent the flux-tube and ambient atmospheres. Right: Inclination of the velocity vector in the flux-tube component. Taken from [5]. As can be seen in Fig. 3, the tubes are inclined upward in the inner penumbra and downward in the mid and outer penumbra. The flow along the tubes is parallel to the magnetic field at all radial distances. The agreement is remarkable, but it hides a serious difficulty: a single flux tube cannot extend across the penumbra with the inclinations of Fig. 3, because it would quickly leave the line forming region (even if the Wilson depression is taken into account). A possible way out of this problem is that the values shown in Fig. 3 do not represent individual tubes, but rather azimuthal averages over short flux tubes whose number density is constant with radial distance (cf. [36]). The flux-tube properties and their radial variation, as derived from Stokes inversions, agree well with those resulting from simulations of moving tubes in the thin tube approximation [31]. The simulations provide a natural explanation for the Evershed flow in terms of a pressure gradient that builds up along the tube as it rises buoyantly from the magnetopause and cools off by radiative losses near the solar surface. The moving tube model explains the motion of bright penumbral grains toward the umbra and the overall morphology of penumbral filaments in continuum images. It also gives convincing arguments why the flux tubes possess more horizontal and weaker fields than the ambient atmosphere, and why the flux tubes return to the solar surface in the mid and outer penumbra (i.e., why their field inclinations are larger than 90o , cf. Fig. 3). The apparent inability of moving tubes to explain the surplus brightness of the penumbra [35] has been used by [41] as an argument to propose the gappy penumbral model. However, the remark made in [35] that dissipation of the kinetic energy of the Evershed flow could account for the penumbral brightness has been overlooked by [41]. Rejecting the idea of hot Evershed flows as the origin of the penumbral brightness cannot be done without 2D or 3D simulations of the evolution of flux tubes including a realistic energy equation and stratified atmospheres. The very existence of flux tubes embedded in a more vertical field has been put into question alleging that such a configuration is not force-free [41]. The imbalance of forces at the top and bottom of the tubes would cause a verti- A Look into the Guts of Sunspots 3 V/Ic [%] 1 2 0 1 -1 0 -2 -1 -2 -3 0.00 0.05 0.10 0.0 0.2 0.4 0.0 0.2 0.4 2 1 V/Ic [%] 275 1 0 -1 0 -2 -1 0.00 0.05 0.10 4 1 V/Ic [%] 2 0 0 -2 -1 -4 -6 -2 0.00 0.05 ∆λ [nm] 0.10 0.0 0.2 ∆λ [nm] 0.4 Fig. 4. Simultaneous spectropolarimetry of AR 10425 with TIP and POLIS at the German VTT of Observatorio del Teide on August 9, 2003. Left: Stokes V profiles of Fe i 630.15 and 630.25 nm. Right: Stokes V profiles of Fe i 1564.8 and 1565.2 nm. The three rows correspond to three different pixels in the limb-side penumbra (θ = 27o ). Filled circles are the observations. Solid lines give the best-fit profiles resulting from an uncombed inversion of the data using the code described in [2]. Adapted from [1]. cal stretching that would eventually destroy the tubes. However, it has been demonstrated [12] that the vertical stretching is limited by buoyancy in convectively stable (subadiabatic) layers. Also, it has been shown that penumbral tubes can be brought into exact force balance if the field within the tube has a small transversal component [8]. Interestingly, the temperature distributions derived from the condition of magnetohydrostatic equilibrium of penumbral tubes produce dark-cored filaments whose properties are very similar to the observed ones [8]. The ability of the uncombed model to explain the existence of dark-cored penumbral filaments has also been demonstrated by means of 2D heat transfer simulations of flux tubes carrying a hot Evershed flow [28]. From a modeling point of view, even the most complex Stokes inversions of penumbral spectra use only two rays to describe the flux tube and the ambient field, which is a very simplistic approximation (see [4] for details). Actually, the two rays represent homogeneous tubes with square cross sections and ambient field lines that do not wrap around the tubes. More sophisticated treatments of the uncombed penumbra are thus desirable for a better interpretation of the observations. Such treatments could remove the small differences between observed and best-fit profiles (Fig. 4). However, one should not expect qualitatively different results, since the uncombed models implemented in current 276 L.R. Bellot Rubio inversion codes already capture the essential physics needed to explain the shapes of visible and infrared lines. 2.2 MISMA penumbral model The MISMA model assumes that the penumbra is formed by optically thin magnetic fibrils a few km in diameter [23, 24]. Each resolution element contains a messy bunch of field lines with random strengths and inclinations that, for an unknown reason, are more or less parallel to the radial direction. The model, implemented in practice as a simple two-component atmosphere, successfully reproduces the asymmetries and NCPs of the Fe i 630.15 and 630.25 nm lines observed in sunspots at a resolution of ∼100 [24]. According to MISMA inversions, downward flows with velocities that often exceed 20 km s−1 exist everywhere in the penumbra [24]. This result is at odds with observations: 0.002 resolution Dopplergrams show no evidence for downflows in the inner and mid penumbra [17]. In addition, the mechanism whereby the small-scale fibrils get organized to produce the large-scale (filamentary) structure of the penumbra remains unknown. This is indeed a serious problem, because negligible azimuthal fluctuations of magnetic field and velocity should be observed when both the number of fibrils per resolution element is large and the fibrils follow the same (random) distribution in different pixels. As a proof of physical consistency, the MISMA deduced from the inversion was shown to satisfy the ∇ · B = 0 condition, unlike simpler one-component models. However, azimuthally averaged atmospheric parameters were used rather than individual values. Since ∇ · B = 0 must be verified locally pixel by pixel, this test does not really demonstrate the validity of the model. It remains to be seen whether MISMAs are able to explain the shapes and NCPs of infrared lines, as well as the existence of dark-cored penumbral filaments. It is also necessary to find reasons why the magnetic fibrils that form the lateral brightenings of dark-cored filaments know of each other so well as to make them move coherently. If MISMAs are the building blocks of the penumbra, regions with zero NCPs will not be detected even at high spatial resolution, because there will always be fibrils interlaced along the LOS. This is perhaps the most important prediction of the MISMA model. 2.3 Gappy penumbral model The gappy model represents a theoretical attempt to explain the existence of dark-cored penumbral filaments and the brightness of the penumbra [41, 30]. It postulates that dark-cored filaments are the signatures of radially oriented, field-free gaps located just below the visible surface of the penumbra. Such gaps would sustain normal convection, thereby providing energy to heat the penumbra. This raises a serious problem, because the existence of vigorous field-free convection plumes reaching the solar surface contradicts the accepted view [40] that the penumbra is deep (as opposed to shallow). A Look into the Guts of Sunspots 277 Another problem is that it is not clear how the model can generate magnetic fields pointing downward in the outer penumbra: the maximum field inclination in a gappy penumbra is 90o , representing horizontal fields. Last, but not least, the model does not offer any explanation for the Evershed flow. It does not even have a suitable place to accommodate horizontal flows, because they must reside where the field is nearly horizontal. Since this happens only in very small volumes just above the gaps, a large fraction of the line forming region would be devoid of flows. The gappy model may be regarded as a limiting case of the uncombed model with zero field strengths in the flux-tube component. The essential difference is that a strong Evershed flow moves along the tube in the uncombed model, whereas in a gappy penumbra not even the field-free regions harbor radial outflows. Thus, an important ingredient for spectral line formation is missing in the model: the discontinuous velocity stratifications produced by confined Evershed motions several km s−1 in magnitude. Gappy models with potential fields do exhibit gradients of field strength, inclination, and azimuth with height [30], but it is unlikely that such gradients can reproduce the multilobed Stokes V profiles and the NCPs of spectral lines without including strong Doppler shifts in an ad hoc manner. Convection in the field-free gaps alone will not produce large NCPs or multi-lobed profiles because (a) it occurs near τ = 1, i.e., far from the line forming region, and (b) the associated velocities will certainly be smaller than 5-6 km s−1 . In summary, although the idea may be appealing, radiative transfer calculations must be performed to demonstrate that the gappy model is able to reproduce the spectropolarimetric properties of the penumbra. Also, heat transfer simulations are required to prove that the field-free gaps would indeed be observed as dark-cored filaments, and that the gaps can heat the penumbra to the required degree. Without these calculations, it seems premature to accept the gappy model as a good representation of sunspot penumbrae. 3 Outlook Currently available models of the penumbra have both strengths and limitations. The difference is that the uncombed model has been extensively confronted with observations, while the MISMA and gappy models still need to pass stringent observational tests to demonstrate their plausibility. Some of the basic claims made by the later models have not yet been confirmed by radiative and/or heat transfer calculations, and hence remain speculative. Further advances in our understanding of the penumbra will come from spectropolarimetric observations at 0.002–0.003. This is the minimum resolution needed to identify the dark cores of penumbral filaments. We would like to measure the vector magnetic fields and velocities of dark-cored filaments not only to distinguish between competing models (which imply different convection modes in the presence of inclined fields), but also to drive holistic MHD 278 L.R. Bellot Rubio simulations of the penumbra. The required observations will be obtained with instruments like the Spectro-Polarimeter [18] aboard HINODE, TIP [13] at GREGOR, IMaX [20] onboard SUNRISE, and VIM [38] aboard Solar Orbiter. References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. C. Beck: PhD Thesis, Albert-Ludwigs Universität, Freiburg (2006) L.R. Bellot Rubio: ASP Conf. Series 307, 301 (2003) L.R. Bellot Rubio: Reviews in Modern Astronomy 17, 21 (2004) L.R. Bellot Rubio: ASP Conf. Series 358, (2006), astro-ph/0601483 L.R. Bellot Rubio, H. Balthasar, M. Collados: A&A 427, 319 (2004) L.R. Bellot Rubio, K. Langhans, R. Schlichenmaier: A&A 443, L7 (2005) L.R. Bellot Rubio, R. Schlichenmaier, A. Tritschler: A&A 453, 1117 (2006) J.M. Borrero: A&A, in preparation (2007) J.M. Borrero, A. Lagg, S.K. Solanki, M. Collados: A&A 436, 333 (2005) J.M. Borrero, S.K. Solanki, A. Lagg, et al.: A&A 450, 383 (2006) J.M. Borrero, S.K. Solanki, L.R. Bellot Rubio, et al.: A&A 422, 1093 (2004) J.M. Borrero, M. Rempel, S.K. Solanki: ASP Conf. Series 358 (2006) M. Collados: this volume (2006) J. Evershed: MNRAS 69, 454 (1909) G.E. Hale: ApJ 28, 315 (1908) M. Landolfi, E. Landi degl’Innocenti: Solar Physics 164, 191 (1996) K. Langhans, G.B. Scharmer, D. Kiselman, et al.: A&A 436, 1087 (2005) B.W. Lites, D.F. Elmore, K.V. Streander: ASP Conf. Ser. 236, 33 (2001) V. Martı́nez Pillet: A&A 361, 734 (2000) V. Martı́nez Pillet, J.A. Bonet, M. Collados, et al.: SPIE 5487, 1152 (2004) S.K. Mathew, A. Lagg, S.K. Solanki, et al.: A&A 410, 695 (2003) D.A.N. Müller, R. Schlichenmaier, G. Fritz, C. Beck: A&A, in press (2007) J. Sánchez Almeida: ApJ 497, 967 (1998) J. Sánchez Almeida: ApJ 622, 1292 (2005) M. Sánchez Cuberes, K.G. Puschmann, E. Wiehr: A&A 440, 345 (2005) T.R. Rimmele: ApJ 445, 511 (1995) T.R. Rimmele, J. Marino: ApJ 646, 593 (2006) B. Ruiz Cobo, L.R. Bellot Rubio: A&A, submitted (2006) G.B. Scharmer, B.V. Gudiksen, D. Kiselman, D., et al.: Nature 420, 151 (2002) G.B. Scharmer, H.C. Spruit: A&A, in press (2006), astro-ph/0609130 R. Schlichenmaier: AN 323, 303 (2002) R. Schlichenmaier, L.R. Bellot Rubio, A. Tritschler: A&A 415, 731 (2004) R. Schlichenmaier, D.A.N. Müller, O. Steiner, M. Stix: A&A 381, L77 (2002) R. Schlichenmaier, W. Schmidt: A&A 358, 1122 (2000) R. Schlichenmaier, S.K. Solanki: A&A 411, 257 (2003) W. Schmidt: ESA-SP 505, 167 (2002) S.K. Solanki: A&A Review 11, 153 (2003) S.K. Solanki, V. Martı́nez Pillet, the VIM Team: ESA-SP 641, in press (2007) S.K. Solanki, C.A.P. Montavon: A&A 275, 283 (1993) S.K. Solanki, H.U. Schmidt: A&A 267, 287 (1993) H.C. Spruit, G.B. Scharmer: A&A 447, 343 (2006) P. Sütterlin, L.R. Bellot Rubio, R. Schlichenmaier: A&A 424, 1049 (2004) A.M. Title, Z.A. Frank, R.A. Shine, et al.: ApJ 403, 780 (1993) C. Westendorp Plaza, J.C. del Toro Iniesta, B. Ruiz Cobo, et al.: Nature 389, 47 (1997) Earth-like Exoplanets. Darwin: Stellar Targets and Precursor Science C. Eiroa1, M. Fridlund2 , L. Kaltennegger2,3, and A. Stankov2 1 2 3 Depto. Fı́sica Teórica, Facultad de Ciencias, Universidad Autónoma de Madrid, Cantoblanco, 28049 Madrid, Spain, carlos.eiroa@uam.es ESTEC/ESA, P.O. Box 299, NL-2200AG Noordwijk, The Netherlands, Malcolm.Fridlund@esa.int Harvard-Smithsonian Center for Astrophysics, MS-20 60 Garden Street, 02138 MA Cambridge, USA, lkaltenegger@cfa.harvard.edu Summary. The ESA Darwin mission will search for extrasolar Earth-like planets within the Habitable Zone of stars, study the physical-chemical properties of their atmospheres, identify potential biosignatures, and carry out comparative planetology. In order to achieve these objectives, a suitable sample of stars with very well known properties have to be selected. In this contribution, the Darwin star catalogue is briefly described, as well as some of the current and future observational efforts aiming to characterize the Darwin targets. 1 Introduction The study of extrasolar planets has become an exciting and active field since the detection a decade ago by M. Mayor and D. Queloz of a Jupiter-like planet orbiting very close to the Sun-twin star 51 P eg [8]. At present, more than 200 extrasolar planets are known, whih that figure increasing at a rate of approximately 1-2 per month (see “The Extrasolar Planets Encyclopaedia” at http://exoplanet.eu). Most of the planets have been discovered using the radial velocity method; in addition, few of them have been discovered by microlensing and by direct imaging. The radial velocity method relies on the Dopler effect produced by the gravitational interaction between the planet and its host star; it determines the minimum mass of the planet because of the uncertainty on the planetary orbit’s inclination. In some favourable cases transits -even secondary ones - have been observed, which resolves the mass ambiguity, and also they allows us to determine the radius of the planet and its density [2, 3]. There is a large diversity of extrasolar planet characteristics and a first, very basic result is that planetary systems are found in many different scenarios. Extrasolar planets found up to now are in some respects similar to the 280 C. Eiroa, M. Fridlund, L. Kaltennegger, and A. Stankov gas giant planets in the Solar System, yet significant differences exist, i.e. the so-called hot-Jupiters orbiting very close to the host stars. The lowest mass planets known up to now are OGLE-05-390Lb, a 5.4 M⊕ planet orbiting a 0.2 M star at a distance of 2.1 AU with a period of 3500 days, and Gliese 876d, a 7.3 M⊕ planet orbiting a 0.32 M star at a distance of 0.02 AU with a period of 1.94 days. Thus, although biased by the detection methods, planetary systems indeed similar to the Solar system and, in particular Earth-like planets, remain undetected. 2 Towards the detection and characterization of Earth-like planets The detection of extrasolar Earths in the Habitable Zones1 of stars, and the characterization of their atmospheres - which further implies the search for biosignatures - constitute the next big challenge in extrasolar planet research. Although microlensing promises to detect Earth-size objects, a shortcoming of the method is the non-repeatability of the lensing event. Thus, photometric transits from space is the most promising method to detect extrasolar Earth-like planets in the near future: CoRoT and Kepler are two such planetary transit space missions (see [9] for a good, overall description of planet detection methods). CoRoT (Convection, Rotation and Transists) is a space mission developed by the French Space Agency (CNES) in collaboration with the European Space Agency (ESA), Belgium, Brazil, Germany and Spain and due to be launched in late 2006. CoRoT has a 27 cm telescope equipped with a widefield CCD camera. More than 100.000 stars will be observed and it is expected that CoRoT will be able to detect the transits of few tens of rocky planets similar to the Earth, but found much closer towards their primary stars. Kepler is a space mission of the American Space Agency (NASA) expected to be launched in 2008. It has a 0.95 m telescope also equipped with a wide field CCD camera which will monitor more than 100.00 stars during the whole 4 year life mission. Kepler is aimed to detect “true” extrasolar Earths, thus orbiting their stars within the Habitable Zone. CoRoT and Kepler will deliver important statistical information about the number and size of extrasolar planets, and specifically about the fraction of stars with Earth-like planets. The detailed characterization of extrasolar Earth-like planets require spectroscopic missions. The space missions Darwin (ESA) and Terrestrial Planet Finder (TPF, NASA) aim at the detection and detailed study of Earth-like planets orbiting within the Habitable Zone of nearby (≤ 25 pc) stars. These missions will either consist of free-flying space infrared interferometers (Darwin, TPF-I) or a large optical space coronagraphic telescope (TPF-C). 1 The Habitable Zone around a star is defined as the zone around a star within which liquid water can be present; it primarily depends on the star luminosity Earth-like Exoplanets. Darwin: Stellar Targets and Precursor Science 281 3 Brief description of Darwin Darwin is based on the assumption that one can spectroscopically characterize the physical and chemical properties of a planetary atmosphere. Darwin (and also TPF) has the unique capability of investigating a broad diversity of planets to understand their formation, evolution and interpreting potential biomarkers. Habitability or biological signatures in extrasolar planets can only be inferred from observations of the reflected or emitted radiation. The direct detection of a planet like the Earth orbiting its host star in the Habitable Zone is a real challenge, since the signal from the planet is ∼ 10−10 − 10−11 (visual range) or ∼ 10−6 − 10−7 (mid-IR range) times the signal from the nearby star. Thus, the selection of an appropriate spectral region to characterize the planets is governed by this high flux ratio and the presence of spectral features indicative of habitability. Darwin will observe in the wavelength region 6 - 20 µm, i.e. the range to detect the thermal emission from the planet, which also is the range where atmospheric terrestrial features as e.g. CO2 , H2 O, CH4 and O3 are found. Darwin consists of several free-flying telescopes building up a nulling interferometer operating in the mid-infrared. Nulling interferometry will cancel the light from the host star to the level required to detect the planetary signal directly. This implies that achromatic phase shifts are applied to the beams collected by the individual telescopes before recombination such that the onaxis light, i.e., the stellar light, is cancelled by destructive interference, while the much weaker planetary light emitted at a small, off axis angle inteferes constructively. 4 The Darwin target star catalogue The key scientific objectives of Darwin are the search for terrestrial extrasolar planets, the detection of potential biosignatures in the planet atmospheres and, in addition, planetology by comparing planet properties, e.g. as a function of stellar age. To achieve these objectives, an input catalogue including appropriate stars has to be prepared. When selecting suitable targets for Darwin, criteria have to be established in order to consider stars for which it is reasonable to assume than an Earthlike planet has formed and, eventually, evolved to the stage where it could in principle host life as we know it. Observational constraints are imposed by the observational method and the Darwin interferometric configuration and architecture. Kaltenegger et al. [6] use Hipparcos as the reference catalogue to make a first selection of potential Darwin targets, since Hipparcos constitutes an homogeneous database, specially for distance estimates. Those authors use the following, basic criteria to the Hipparcos stars in order to obtain a preliminary Darwin target list: i) stars located at distances less than 25 pc; ii) stars located in a cone of aperture ±45o around the ecliptic; iii) main-sequence 282 C. Eiroa, M. Fridlund, L. Kaltennegger, and A. Stankov (luminosity classes V or IV/V) F GK spectral type stars. Those criteria take into account both the instrumental constraints and the assumptions about the presence of Earth-like planets and their habitability (the final catalogue will also include some M stars). The number of F GK Hipparcos stars which satisfy the criteria are 74 F , 142 G and 288 K stars. Fig. 1 is a plot of the position of the F GK stars in the Darwin preliminary catalogue in equatorial coordinates; it can be appreciated that the stars are rather uniformly distributed in the sky area accesible for Darwin observations. Fig. 2 is the distance histogram of the stars; as expected, most of them are located at distances between 20 and 25 pc. Considering Hipparcos completeness, the sample should be fairly complete for the F and G stars, but for late type K stars it is most likely uncomplete. Fig. 1. Sky distribution of the preliminary Darwin F GK stars in equatorial coordinates. The modulation is due to the ±45◦ ecliptic cone. Recent ESA studies show that the interferometer configuration can significantly simplify with respect to the “classical” one [5]. This could eventually mean that the whole sky would practically be accesible for the Darwin search of extrasolar Earths, as well as it would be possible to increase the size of the telescopes’ mirrors, increasing in this way the distance of the stars which could be observed by Darwin. Nonetheless, given the required very long exposure times for the detection plus the characterization of the planetary atmospheres, the core programme will in any case consist of stars closer than 25 pc. One of the most relevant observational constraints for Darwin is the multiplicity/binarity nature of the targets. Its influence is twofold: i) a faint object within the field of view of the interferometer or in the immediate surroundings can prevent Darwin to obtain a clean planetary signal, since the interferometric null is compromised by the multiplicity nature of the star. This effect can be present in both physical multiple systems or projected field stars. ii) The existence of a physical companion can influence the proper existence of an Earth-like Exoplanets. Darwin: Stellar Targets and Precursor Science 283 Fig. 2. Distance histogram of the stars in the Darwin preliminary target list Earth-like planet in the stellar Habitable Zone. After consulting a number of public catalogues on stellar multiplicity, spectroscopic binaries, and eclipsing binaries, Kaltenegger et al. ([6]) identify 52 F stars, 106 G stars, and 226 K stars without known companions at angular distances ≤ 5 arcsec. Some of the Darwin stars are known to host extrasolar giant planets. The presence of a giant planet, even hot-Jupiters, does not necesarily exclude terrestrial planets in the stellar Habitable Zone, e. g. [10]. Thus, at least some of the stars with giant planets can be consider good Darwin candidates, although nulling interferometry simulations are needed to asses the detectability of Earth-like planets in planetary systems similar to those known with giant planets. 5 Observational precursor science A knowledge as deep as possible of the properties of stars potentially harbouring Earth-like planets is required, since planet atmospheres are largely influenced by the radiation received from the star [11]. The characterization of the host stars has to be achieved by the analysis of astrophysical data in existing data archives and catalogues, as well as by means of an observational roadmap to cover the areas where existing data are insufficient or inadequate. Parallel, theoretical and modeling efforts have to be developed in aspects like planetary formation within protostellar disks, planetesimal formation, formation of giant and telluric planets, orbital stability, planetary migration, evolution of planet atmospheres, etc. This research will allow us to establish optimization criteria to be applied to the preliminary Darwin target list, identifying those stars that are most likely to harbour Earth-like planets and that do not offer an hostile environment for the development of life. 284 C. Eiroa, M. Fridlund, L. Kaltennegger, and A. Stankov From the observational point of view, properties of the stars themselves and of their immediate environments are to be known in detail. A (nonexhaustive) list of such properties are: • Host stars: fundamental astrophysical parameters like temperature, luminosity, radius, rotation, mass, gravity, age, metallicity; photometric behaviour, induced variablity due to star spots, pulsations; chromospheric activity, flares, magnetic fields, stellar winds, etc. • Stellar environment: exo-zodiacal disks and Kuipert belts; faint and very faint (planets, brown dwarfs) physical and non-physical companions, distances and proper motions of companions; physical membership to stellar associations or groups, etc. We have already started a study of the Darwin stars [4] by systematically consulting a variety of catalogues, public archives and papers with photometric and spectroscopic information, as well as estimates of stellar properties and the environment. In addition, we are developing a Darwin archive to collect all this information as well as future data, based on the Virtual Observatory standards and procedures [12], which will be open to external use in the next future. As an example of our work, Fig. 3 shows the HR diagram of the stars in the preliminary Darwin target list. The bolometric luminosity of individual stars has been estimated from published photometry, bolometric corrections according to the published spectral types, and Hipparcos parallaxes, while the effective temperature has been assigned according to the spectral types, and also estimated from photometry. The plot shows that some stars do not behave as expected from main sequence stars, which clearly demonstrates the Fig. 3. HR diagram of the F GK stars in the preliminary Darwin target list. Some stars are not located along the main sequence, which means that they require a critical analysis before being selected as good Darwin candidates. Earth-like Exoplanets. Darwin: Stellar Targets and Precursor Science 285 need of a better characterization, even in such a basic astrophysical aspect as the HR diagram. There are already some running, observing programs with the aim of contributing to the precursor science required by Darwin. High resolution spectroscopic observations of Darwin stars have been carried out at La Palma and Calar Alto observatories; first results are shown in these Proceedings [7]. Spitzer is observing in the mid-/FIR some TPF stars, achieving flux level of around 100 solar zodies [1]. In addition, there are plans to study extrasolar Kuipert-belt structures with Herschel. Further, there are also plans to carry out high spatial resolution observations with high flux contrast using interferometric and AO systems, in order to study the environments of the stars. These are just few examples of some activities which are being undertaken in order to make it posible that Darwin finds and characterizes Earths around nearby stars. References 1. 2. 3. 4. 5. 6. 7. 8. 9. C. A. Beichmann, A. Tanner, G. Bryden, et al.: ApJ 639, 1166 (2006) C. Charbonneau, T.M. Brown, D.W. Latham, M. Mayor: ApJ 529, L45 (2000) D. Charbonneau, L.E. Allen,, S.T. Megeath, et al.: ApJ 626, 523 (2005) C. Eiroa et al.: in preparation (2006) ESA-SCI 12: Darwin. The infrared Space Interferometer (2000) L. Kaltenegger, C. Eiroa, A. Stankov, C.V.M. Fridlund: in preparation (2006) J. Maldonado et al.: these Proceedings (2006) M. Mayor, D. Queloz: Nature 378, 355 (1995) M. Perryman, O. Hinault: Extra-Solar Planets, ESA-ESO Working Groups, Report No. 1, March 2005 10. S. N. Raymond, R. Barnes, N. A. Kaib, ApJ 644, 1223 (2006) 11. F. Selsis: in Earth-like planets and moons. Proc. 36th ESLAB Symposium, p. 251 (2002) 12. E. Solano, C. Eiroa,, et al.: in preparation (2006) How the comet 9P/Tempel 1 has behaved before, during and after the Deep Impact event L.M. Lara1 , H. Boehnhardt2 and P.J. Gutiérrez1 1 2 Instituto de Astrofı́sica de Andalucı́a, CSIC, Camino Bajo de Huétor 50, 18008 Granada, Spain, lara@iaa.es, pedroj@iaa.es Max-Planck Institut fuer Sonnensystemforschung, Max-Planck Str. 2, Katlenburg-Lindau, D37189 Germany, boenhardt@mps.mpg.de Summary. Comet 9P/Tempel 1, the target of the Deep Impact Mission, has been monitored for 7 months and a half aiming at its characterization before, during and after the impact experiment. This characterization in each phase comprises the (i) determination of the rotation axis, (ii) evolution of the gas and (approximate) dust production rates, (iii) analysis of the gas and dust radial profiles, (iv) study of the dust colour as a function of the heliocentric distance and projected cometocentric distance, and (v) searching for a new long-lasting morphological structure in the coma due to the DI experiment. 1 Introduction In January 2005 NASA’s Deep Impact (DI) spacecraft was launched to perform a cratering experiment at Comet 9P/Tempel 1 on 4 July 2005: a 362 kg impactor hit the cometary nucleus at 10.2 km/s speed to excavate a crater and to initiate new activity of the nucleus [2], [1]. The scientific goals of this unique experiment are, apart from studying crater physics, to characterize the nucleus P/Tempel 1 as much as possible, as representative of the primordial bodies from the formation period of the planetary system. Due to the very limited instruments on board the fly-by spacecraft (two cameras and a spectrometer) that can follow the impact for 800 s, a significant science contribution is expected from Earth-based observations. In fact, the mission was designed to have much of the mission-critical science done from Earth-based telescopes, and an overview of the scientific conclusions and collective observations from the world-wide campaign has been presented elsewhere [12]. Calar Alto Observatory (CSIC-MPG) and Sierra Nevada Observatory have participated in the world-wide campaign carried out to study the comet 9P/Tempel 1 at preimpact, impact and post-impact phases between early January 2005 and July 12, 2005. The results of these campaigns, aiming at a characterization of the comet behaviour before, during and after the DI event, are presented here. 288 L.M. Lara, H. Boehnhardt and P.J. Gutiérrez 2 Observations and data reduction Comet 9P/Tempel 1 was monitored from the Calar Alto Observatory (CSICMPG near Almeria, Spain) and Sierra Nevada Observatory (CSIC, near Granada, Spain) since Jan. 2005, using the instruments CAFOS (in imaging and spectroscopic modes) and BUSCA mounted at the 2.2 m telescope at Calar Alto, and Versarray CCD at the 1.5 m telescope at the Sierra Nevada Observatory. The comet images, and overheads to render calibration, were acquired with Johnson B,V,R and I broadband filters, typicaly every 3 to 5 days from Calar Alto Observatory and during selected periods from the Sierra Nevada Observatory. The spectra were taken once every month and they covered observable spectral ranges between 3200 and 8800 A with a wavelength scale of 4.75 A per pixel, and between 2800 A and 1.0 µm with a wavelength scale of 9.7 Å per pixel, respectively. The slit of the spectrograph was orientated in north-south direction, giving dust and gas profiles at different cross-cuts through the coma, depending on the comet position in the sky. For absolute calibration, observations of appropriate spectrophotometric standard star were acquired too. Beside these periodic observations, two in situ runs of 5 and 12 days were carried out in mid-April and in July. All comet observations were done with telescope tracking at the comet’s proper motion. With the exception of the 5 night run in April 2005 and the 12 nights in July, all observations were done in service mode at both observatories. Details on the image and spectral reduction and calibration can be found in [9], [10], [11]. During the time the comet has been observed, it has moved inbound from (rh , ∆) = (2.251, 1.878) AU to its perihelion on July 5 at (rh , ∆) = (1.506, 0.902) AU. The perigee at 0.711 AU took place on May 4 when the heliocentric distance was rh = 1.628 AU. The position angle of the Sun-comet vector has ranged from 290.2◦ to 105.9◦, whereas the Sun-comet-observer angle varied from 25.49◦ to a minimum of 11.09◦ . 3 Results on coma morphology 3.1 Pre-impact phase For the enhancement of morphological structures in the coma calibrated R and I filter images (for examples see Fig. 1) are processed maily by applying the Adaptive Laplace filtering [5] (see Fig. 2). The sequence depicts the evolution of coma structures in the comet between early January 2005 until a month before impact. The porcupine coma structure with only minor changes of the near-nucleus position angles with time suggests the interpretation as an embedded fan coma [14]. The borderlines of these projected fans are the straight or curved features seen in our enhanced images (however, not every cone may produce two sharp border lines [13]). The number of straight jets found in our images suggests the presence of at least 3 or 4 very active regions on the nucleus, depending on the assumed association of these jets with 9P/Tempel 1 and the Deep Impact event 289 borderlines of coma fans. A first and conclusive assessment of the slow drift of position angles observed supports a rotation axis orientation close to the angular momentum vector of the orbital motion of the comet, as it has been confirmed after the DI event from an exhaustive analysis of the experiment results [16]. 3.2 Impact and post-impact phases The study of dust dynamics is done from the broadband images (R and I Johnson filters). The images show - after processing as described in [6] and [10]- the projected geometry of dust structures in the cometary coma. Figure 3 provides a view of the temporal evolution of the coma structures from June 26 to July 5, 2005. Before the impact, the only existing structures are those already described by [10], whereas ∼ 15 hrs and 40 hrs after the impact the dust cloud produced by the DI experiment is clearly seen, as well as its expansion in the south-west direction. More concisely, the leading edge of the cloud ejecta extends up to ∼ 13 000 km in position angles 125◦ to 350◦ whereas forty hours after impact, i.e. July 5, the expanding dust cloud forms a shell which has noticeably changed its shape due to the push of the solar radiation pressure forcing the particles into the tail. It reaches ρ ∼ 23 000 km in the sunward direction (PA of the Sun is ∼ 297◦ counted north over east). At that time, the coma structures existing before the impact, are mostly hidden by the ejecta plume. However, they became clearly visible again when the ejecta cloud had expanded and attenuated over the following days. By dividing the images obtained on July 4 and 5 by those on July 3, the expansion velocity, projected on the plane of the sky at PA=(225 ± 20)◦ , of the leading edge of the ejecta cloud can be computed giving rise to ∼ 230 m/s and ∼ 152 m/s, 15 and 40 hours after impact. The ejecta cloud is still visible in our Calar Alto images of July 6. No new long-lived jet or fan, as a consequence of the impact crater, is detected neither in our Calar Alto images nor reported from other observations around the world (to our knowledge). Let us note that the Calar Alto monitoring lasted for 8 days, or ∼ 5 nucleus rotations. Signatures of the ejecta cloud were still seen in imaging observations on July 7 [12]. 4 Results on gas and dust activity 4.1 Pre-impact phase Estimates of the dust production in comets are usually made by means of the parameter A(θ)f ρ [4] as a function of the projected cometocentric distance ρ. To compute this parameter we make use of the comet images acquired in R Johnson filter (which contains little or no gas contamination, thus representing Sun light scattered by the dust grains). The Af ρ value varies with heliocentric distance as rh−6.71 and it ranges from 55.8 cm at rh = 2.21 AU when the 290 L.M. Lara, H. Boehnhardt and P.J. Gutiérrez monitoring started, to a maximum value of 287.7 cm at rh = 1.71 AU on April 14, 2005, long before the perihelion. Slightly enhanced Af ρ (above the rh−6.71 curve) was observed from mid-February until the end of March. Between midApril to mid-May (i.e. 60 to 80 days before perihelion) the comet peaked on dust activity. From May 11 to June 14 the comet has experienced a continuous decrease in the dust production rate with some sporadic events of increased activity. The spectroscopic measurements, as well as the R and I images have allowed us to determine the average dust colour, or gradient reflectivity, within the coma as S 0 ∼ 20 − 30%/100 nm, with no variations either as function of the projected cometocentric distance or with heliocentric distance while the comet has moved inbound. The CN, C2 and C3 production rates, Q have been derived from the spectroscopic measurements acquired every 0.1-0.2 AU during the long-term preimpact monitoring by making use of the Haser modeling [7] and customary gas expansion velocity and lifetime [3]. Clear gas emissions were detected during the 5 days run in April, when the log Q for CN and C2 was in the order of (1 − 3) × 1024 s−1 , whereas for C3 the production rate was an order of magnitude lower. 4.2 Impact and post-impact phases The approximate dust production rate of 9P/Tempel 1 was monitored from June 24 to July 12. Beside the increase in Af ρ produced by the Deep Impact experiment on July 4 (whose aftermath still measurable on July 5), we note some sporadic increases in the dust production rate on Jun 24, June 29 (likely connected to a natural outburst reported in mid-IR (Wooden et al. private communication) observations which increased the comet brightness in a ∼ 20%. In a very approximate way, the Af ρ measured on July 4 can be used to estimate the total mass of dust produced by the impact itself. For this, we have subtracted the Af ρ value measured on July 3 in an aperture of 5 000 km from the one on July 4. This gives rise to Af ρimpact = 76 cm, equivalent to ∼ 14.5 hrs of pre-impact regular activity. Dust color maps have been computed as the normalized reflectivity gradient S 0 in %/100 nm. On every date, these maps do not show variations of the grain properties (size and/or composition) in the coma of 9P/Tempel 1 in the inner coma excluding July 04. At about fifteen hours after the impact, the dust color within the ejecta plume is bluer than in the rest of the coma up to projected distances of ρ ∼ 15 000 km. However, on July 05, 40 hours after impact, there dust coma shows a weak trend toward an overall blueing at ρ ≤ 15 000 km, whereas on July 07, S 0 returns to a value of ∼ 20%/100 nm. Figure 4 displays the dust color variations in 2D. It can be seen that the reflectivity gradient is lower in those directions that are populated by the grains ejected by the impact, i.e. the dust color is bluer than in adjacent regions, meaning that either there is an overpopulation of submicrometer to micrometer dust grains [15], [8], [1], [12] or they are more refractive in the blue range than in the red one. 9P/Tempel 1 and the Deep Impact event 291 Gas profiles (emission flux and column density) for CN, C2 and C3 vs projected cometocentric distance ρ have been derived from the spectroscopic observations between July 2 and 8, (excluding July 7 due to passing clouds). The production rates of these species remained rather stable during those days, with values around (1.0 − 1.5) × 1025 s−1 for CN and C2 and an order of magnitude lower for C3 , excluding July 04.917 UT. At that date, the aftermath of the DI event is well seen in every radial profile in the north-south direction as an increased number of column density at those directions at distances lower than ∼ 30 000 km (beyond this distance, the spectra have a low S/N). The total number of CN, C2 and C3 molecules produced by the DI collision, integrated in a circular aperture of radius 30 000 km, is 2.13 × 1029, 2.07 × 1029 and 1.52 × 1028 , respectively, or similarly 12 200 kg of CN, 19 000 kg of C2 and 7 900 kg of C3 . In a general way, the nuclear activity, in terms of gas and dust (parameterized by Af ρ) production rates, returned to pre-impact levels on July 06, 2005. Fig. 1. Isophote images of Comet 9P/Tempel 1. The sequence shows the evolution of the dust coma from January to mid June 2005 (dates are given in the figure). North is up and East to the left. The field of view is 2.25×2.25 arcmin. The brightness peak in the coma (equivalent to the nucleus position) is placed in the center of the field of view. 5 Overview Our monitoring of the comet 9P/Tempel 1, the target of the Deep Impact Mission, has been carried out from January to mid-July 2005 aiming at its characterization before, during and after the impact experiment. Before the impact, i.e. from January to end of June, the coma has gone through a slow morphological evolution from a wide structure in the southwestern quadrant in mid-February to a porcupine pattern in mid-April and 292 L.M. Lara, H. Boehnhardt and P.J. Gutiérrez Fig. 2. Laplace filtered images of the coma structures of Comet 9P/Tempel 1. The sequence shows the Laplace filtered version of the images in Fig. 1. Laplace filtering enhances shortscale brightness gradients in the images, while at the same time suppressing the longscale ones. The identified coma structures listed in Table 2 of [10] are labeled in the lower left sub-image. Orientation and field of view as in Fig. 1. Fig. 3. Laplace filtered images of the coma structures of comet 9P/Tempel 1 as imaged in the R Johnson filter on June 26, 29, July 04 and 05 from top left to bottom right. North is up and East to the left. The field of view is 60 000 × 60 000 km and the nucleus is at the center of the FOV. up to seven features identified in June. In addition to this evolution, an arclet in the western coma hemisphere was first detected on June 14, related to an outburst event, and afterwards confirmed by the Hubble Space Telescope. Interpretation of these features and their evolution seems to indicate the presence of at least 3 or 4 very active regions on the nucleus, consistent with the rotation axis being close to the angular momentum vector of the orbital motion of the comet. The value of Af ρ varies with heliocentric distance as rh−6.71 slightly enhanced Af ρ (above the rh−6.71 curve) was observed from mid-February until 9P/Tempel 1 and the Deep Impact event 293 Fig. 4. Two dimensional color map of the dust in the coma of 9P/Tempel 1 on July 4, i.e. ∼ 15 hrs after impact. North is up and East to the left. The look-up table is linear between 0 and 30%. The field of view is 60 000 × 60 000 km at the comet distance centered at the nucleus. the end of March, when fan-shaped structures appeared in the coma for the first time. Somewhere between mid-April to mid-May (i.e. 80 to 60 days before perihelion), the comet peaked in dust activity. In terms of gas production rates, CN, C2 and C3 have been obtained at rh ∼ 1.7, 1.60 and 1.51 AU, being slightly below those derived from previous passages. Abundance ratios indicate that 9P/Tempel 1 is classified as a typical comet in terms of C2 abundance. The surface brightness profiles of the continuum, either azimuthally averaged profiles from the broadband images or in north-south direction from the long-slit spectra can be well fit with −1.9 ≤ m ≤ 1.14 in log B− log ρ representation. Steeper slopes are obtained at larger rh which might be related to variable dust size distribution with distance from the nucleus due to the radiation pressure dynamics and/or physical processing of the dust grains (sublimation, fragmentation). Normalized color of the dust inside the coma in the north–south direction is measured to be S 0 ∼ 20 − 30%100 nm. The Deep Impact event took place on July 04.226 UT 2005. Fifteen hours after the impact, the ejecta cloud extends over ∼ 240◦ in position angle (PA) with symmetry axis at PA∼ 225◦. The effect of the solar radiation pressure is already visible as a slight deviation from a fully symmetric plume and the ejecta dust is already feeding the tail. The exhaustive analysis of the broadband images has revealed that no new long lasting coma structure is produced by the impact. The structures existing in the coma before the event are recovered after the ejecta plume has moved out. The maximum projected expansion velocity of the ejecta dust results into ∼ 230 and ∼ 150 m/s 15 and 40 hours after impact, respectively. Surface brightness profiles of the continuum, either azimuthally averaged profiles from the broadband images or in the north-south direction from the 294 L.M. Lara, H. Boehnhardt and P.J. Gutiérrez long-slit spectra can be generally well fit with slope m of −0.94 ≤ m ≤ −1.49 in log B− log ρ representation. A few exceptions occur on July 2 and 8-10 when much flatter continuum profiles are detected possibly related to fragmentation processes and to the reported outbursts occurring around those dates. Normalized color S 0 of the dust inside the coma does not show spatial variations excluding July 04.875 UT, our first observation after the impact. At that time, the dust inside the ejecta plume is undoubtedly bluer than the surrounding coma (8.2 ± 0.4%/100 nm versus 14.5 ± 0.8%/100 nm). The dust colour in the whole coma returns to a ∼ 20%/100 nm on July 07.875 UT, the same value measured prior to the projectile impact for a considerable period of time. A lower limit to the mass in the ejecta can be given from our optical observations resulting into 1.2 × 106 kg, which represents about 14 hours of quiet (i.e. steady state) pre-impact activity. The value of Af ρ is remarkably variable during the 18 days monitoring as several outbursts took place, beside the one induced by the DI experiment. Apart from outburst periods, Af ρ ∼ 110 − 120 cm. The gas activity represented by the CN, C2 and C3 production rates (Q), are relatively constant from July 1 to 6 excluding the immediate post-impact period on July 04. The number of molecules of CN, C2 and C3 produced by the DI were equal to 2.13 × 1029 , 2.07 × 1029 and 1.52 × 1028 . The amount of their potential parent species detected at other wavelengths seems to indicate that a large fraction of the daughter species measured 15 hours after the impact, might originate from the ejected dust grains. References 1. A’Hearn, M.F., et al.: Science 310, 265 (2005) 2. A’Hearn, M.F., Belton, M.J., Delamere, A. & Blume, W.H.: Space Sci.Rev. 117, 1 (2005). 3. A’Hearn, M.F., Millis, R.L. Schleicher, et al.: Icarus 118, 223 (1995) 4. A’Hearn, M.F., Schleicher, D.G., Millis, R.L., et al.: Astron. J. 89, 579 (1984). 5. Boehnhardt H., Birkle K.: A&A Suppl. 107, 101 (1994). 6. Boehnhardt H., Birkle K., Fiedler A. et al.: EMP 78, 179 (1999). 7. Haser, L.: Bull. Cl. Sci. Acad. R. Belg., 43, 740, (1957). 8. Harker, D.E., Woodward, C.E. and Wooden, D.H.: Science 310, 278 (2005). 9. Lara, L.M., Rodrigo, R., Tozzi, G. P, Boehnhardt, H., Leisy, P.: A&A 423, 1169 (2004). 10. Lara, L.M., Boehnhardt, Gredel, R., Gutiérrez, P.J., Ortiz, J.L., Vidal-Nuñez, M.J., Rodrigo, R.: A&A 445, 1151 (2006) 11. Lara, L.M., Boehnhardt, Gredel, R., Gutiérrez, P.J., Vidal-Nuñez, M.J., Rodrigo, R.: A&A, in press. (2006). 12. Meech K.J., Ageorges, N., A’Hearn, M.F., et al.: Science, 310, 265 (2005). 13. Sekanina Z.: ESA SP-278, 315 (1987). 14. Sekanina Z., Boehnhardt H.: EMP 78, 313 (1999). 15. Sugita, S., et al.: Science 310, 274 (2005) 16. Thomas P. et al: Icarus, in press, (2006). Heliospheric energetic particle variability over the solar cycle D. Lario The Johns Hopkins University, Applied Physics Laboratory, 11100 Johns Hopkins Rd. Laurel MD 20723-6099, USA, david.lario@jhuapl.edu Summary. The energetic particle contents of the heliosphere change from solar maximum to solar minimum. The ultimate responsible for those variations is our changing Sun. Its changes are reflected in the dynamics of the large-scale structure of the heliosphere, the solar output of energetic particles, and definitively, in the origin, intensity, energy and composition of the population of energetic particles observed by spacecraft and Earth-based detectors. The stable and regular pattern of recurrent energetic particle events observed in association with corotating interaction regions (CIRs) during solar minimum is replaced by the frequent observation of solar energetic particle (SEP) events associated with either solar flares and/or fast coronal mass ejections (CMEs) during solar maximum. The higher frequency of CMEs and transient events during solar maximum results in both a global filling of the inner heliosphere (< ∼ 10 AU) with low-energy particles and a more complex dynamic heliosphere that hinders the penetration of galactic cosmic rays (GCRs). The composition of the low-energy (0.04-1.0 MeV/nucleon) ion population evolves < > from ratios H/He< ∼ 0.20, Fe/O∼ 0.15 and C/O∼ 0.75 during solar minimum, to ratios < > > H/He∼ 0.5, Fe/O∼ 0.25 and C/O∼ 0.5 during solar maximum. 1 Heliospheric Energetic Particle Sources The sources of energetic particles in the heliosphere are diverse, depending on the phase of the solar cycle and the energy of the particles. Galactic cosmic rays (GCRs) originated in interstellar space dominate the proton intensities above about 200 MeV and their intensity is modulated by the solar activity [17]. Below ∼100 MeV the proton intensity averaged over a solar cycle is mainly dominated by events of solar origin. Energetic ions associated with the interaction between fast and slow solar wind streams in interplanetary space (i.e. corotating interaction regions, CIRs) can also be observed at energies as high as several MeV/nucleon at all latitudes and predominantly during solar minimum [19]. During quiet times it is also possible to observe energetic ions accelerated presumably close to the heliospheric termination shock at energies as high as 100 MeV/nucleon (i.e., anomalous cosmic rays, ACRs) [3]. 296 D. Lario The energetic electron population in the heliosphere also changes over the solar cycle. At energies above ∼1 GeV the electron intensity at 1 AU is dominated by galactic cosmic-ray electrons which continuously penetrate the heliosphere. From 3 MeV to 1 GeV galactic cosmic-ray electrons are still observed but their fluxes are attenuated and modulated by solar activity [6]. At solar quiet times, most of the electrons observed in the range from a few hundred keV to a few MeV are of Jovian origin [5]. At lower energies, ∼50 keV, corotating interaction regions accelerate electrons [22]. During the active phases of the solar cycle the emission of solar electrons exhibits a considerable increase. Those electrons are observed at 1 AU in the form of transient events from a few keV to an upper limit of about 100 MeV [14]. The contribution of each one of these sources (i.e., galactic, solar and interplanetary) to the heliospheric energetic particle population changes over the solar cycle. These changes are ultimately determined by our changing Sun and the variations it produces in the large-scale structure of the heliosphere. 2 Variations Over the Solar Cycle The top panel of Figure 1 shows 10-day averages of 0.50-0.96 MeV proton intensities as measured by the Charged Particle Measurement Experiment (CPME) on board the Earth-orbiting Interplanetary Monitoring Platform 8 (IMP-8) [20] from 26 Oct 1973 to 24 Oct 2001 (point where IMP-8 observations were discontinued). The gray traces show 10-day averages of 0.587-1.06 MeV proton intensities measured by the Electron Proton Alpha Particle Monitor (EPAM) on board the Advanced Composition Explorer at the Sun-Earth L1 point [7] from 30 Aug 1997 to present (Oct 2006). The mid-panel of Figure 1 shows daily averages of GCR intensities as measured by the Climax Neutron Monitor in Colorado with a cut-off rigidity of 3 GV. The bottom panel of Figure 1 shows the monthly sunspot number. Figure 1 provides a complete perspective of the effects that solar cycle variations produce on several energetic particle populations. The low-energy proton intensities show an oscillating trace modulated by the sunspot number. Superimposed on this global trend there is an abundance of relatively shortlived particle flux increases (from hours to several days) that are sporadic transient events of solar origin or associated with recurrent CIRs. Noteworthy is the fact that the intensity minima are sustained at a higher level during the active phase of the solar cycles and only return to instrumental level when the solar activity is minimum. This behavior is the result of multiple and frequent particle injections from the Sun that produce the global filling of the inner heliosphere. Occasionally, during periods of sustained high solar activity, the inner heliosphere may act as a reservoir of low-energy particles when large spherical volumes remain filled with energetic electrons, protons and heavy ions for long (>10 days) periods [18][12]. Heliospheric energetic particle variability 297 Fig. 1. Top panel. 10-day averages of the 0.50-0.96 MeV proton intensities as measured by IMP-8/CPME (black trace) and 0.587-1.06 MeV as measured by ACE/EPAM (gray trace). Mid-panel. Daily averages of cosmic ray intensities as measured by the Climax Neutron Monitor. Bottom panel. Monthly (gray traces) and monthly smoothed (black traces) sunspot number for the period Oct 1973-Dec 2006. The composition of the low-energy (0.04-1.0 MeV/n) ion population over the solar cycle is also highly dynamic [4]. During periods of increased solar activity, the suprathermal ion population is dominated by ions accelerated in association with solar events (either solar flares and/or shock waves driven by CMEs): the 0.5-1.0 MeV/n H/He is usually above 50 [13], whereas the 0.08> 0.16 MeV/n C/O and Fe/O ratios are < ∼ 0.5 and ∼ 0.25, respectively [4]. During solar minimum conditions, the suprathermal ion population is more similar to what is usually observed in the solar wind with possible contributions of pickup ions: the 0.5-1.0 MeV/n H/He in CIR events is about ∼17 [13], whereas the 0.08-0.16 MeV/n C/O and Fe/O ratios are ∼0.75 and ∼0.1, respectively [4]. The low values of H/He and Fe/O together with the high values of C/O during solar minimum conditions have been interpreted as a consequence of the acceleration of solar wind and pick-up ions in CIRs [15]. During solar maximum, the larger number of transient events richer in H and Fe, together with the weakening of fast-slow solar wind interactions, contribute to form a suprathermal background with high H/He and Fe/O ratios. 298 D. Lario The mid-panel of Figure 1 shows the count rate of GCRs. This count rate varies inversely with the sunspot number. This intensity curve also shows the ∼22 year cycle with alternative maxima, being flat-topped in the minima of solar cycles 20 and 22, and peaked at the minimum of solar cycle 21. This behavior has been reproduced by models of GCR modulation based on the observed reversal of the Sun’s magnetic field polarity every ∼11 years and curvature and gradient drifts of the particles in the large-scale heliospheric magnetic field (HMF) ([8] and references therein). During solar minimum, when the large-scale HMF is relatively ordered, gradient and curvature particle drifts dominate the cosmic ray modulation. During solar maximum, the increase of the HMF strength, the more complex structure of the heliosphere (with a highly tilted heliospheric current sheet) and the higher frequency of CMEs and shocks propagating outward from the Sun have been suggested as possible causes of the reduced GCR intensities ([2] and references therein). Simunac and Armstrong [24] analyzed the 0.39-440 MeV proton energy spectra over the solar cycles 21, 22 and part of 23, and concluded that during periods of solar minimum the falling energy spectra show upturns between 100 MeV and 345 MeV due to the presence of both GCRs that typically peak between 1000 and 2000 MeV and ACRs that contribute to the fluxes between 20 and 100 MeV. The slope of the spectra at energies below ∼100 MeV is steeper during solar minimum than solar maximum, indicating a larger contribution of higher-energy particles of solar origin (<100 MeV) during solar maximum years. The GCR and ACR contributions are not observed during solar maximum because of the larger SEP contribution at high energies [24]. 3 Occurrence of Solar Energetic Particle Events It is well-known that the occurrence rate of SEP events is higher during the active periods of the solar cycle. In order to determine the number of events occurring within each solar cycle, it is necessary to establish standard criteria to identify each single event. A list of solar proton events that covers from Apr 1974 to present (Oct 2006) has been assembled by the National Oceanic and Atmospheric Administration (NOAA) at umbra.nascom.nasa.gov/SEP/seps.html. A proton event is defined as an episode with >10 MeV proton intensities above 10 particles (cm2 sec ster)−1 (i.e., >10 p.f.u.). Note that this episode may include several injections of particles from the Sun or traveling CME-driven shocks. Figure 2a shows the annual frequency of episodes with >10 MeV proton intensities above 10 p.f.u. together with the monthly sunspot number. Although these episodes may occur at any time over the solar cycle, more events occur during the maximum of solar activity (within around three years of the maximum of the solar sunspot cycle) than during the remaining portion of the solar cycle. Solar cycle 21 was an exception to this trend since events were observed throughout the cycle. Heliospheric energetic particle variability 299 The intensity of the events and the total fluence of particles during solar cycle 21 were less than in the other studied solar cycles [21]. Fig. 2. Monthly smoothed sunspot number (thin line), monthly sunspot number (gray line), and annual frequency of episodes with >10 MeV proton intensities above 10 p.f.u. (a) and annual frequency of GLEs (b). Shea and Smart [21] assembled a list of significant proton events from May 1954 (start of cycle 19) to May 2001 identifying each unique solar proton injection as a discrete event. Thus, each event in an episode of solar proton events that may be associated with the same active solar region as it traverses the solar disk is counted as a separate event. They found that while the distribution of events in time differs from cycle to cycle, the total number of events during each cycle (as per their definition) is remarkably constant (with an average number of 75 events per solar cycle). Significant solar proton events (as per their definition) can occur at almost any time of the solar cycle; and approximately 16% of the events for each solar cycle (cycles 19-22) contain relativistic solar protons as recorded by ground-based neutron monitors. These events are termed ground-level enhancements (GLEs). Figure 2b shows the annual frequency of GLEs together with the monthly sunspot number. As in the case of the episodes with >10 MeV protons above 10 p.f.u., GLEs can be observed at any time during a solar cycle (such as in Nov 1997 and Jan 2005 in the rising and declining phases of solar cycle 23, respectively). GLEs are more frequent close to solar maximum. Solar cycle 21 (and perhaps solar cycle 23) present a more uniform distribution of GLEs. 4 Ulysses: The Heliosphere in Four Dimensions The Ulysses mission (launched on 6 Oct 1990) has already completed two orbits over the poles of the Sun under completely different solar activity conditions. The spacecraft’s unique orbit, almost perpendicular to the ecliptic 300 D. Lario plane, and the broad range of energetic particle measurements being made, make it ideal for studying the energetic particle contents of the heliosphere in four dimensions: space and time. Figure 3 shows, from top to bottom, daily averages of the 38-53 keV electron fluxes as measured by the Heliosphere Instrument for Spectra, Composition and Anisotropy at Low Energies (HI-SCALE) [9] (panels a), the 1.8-4.7 MeV ion intensity as measured by HI-SCALE (panels b), the 71-94 MeV proton intensity as measured by the Ulysses Cosmic Ray and Solar Particle Investigation (COSPIN) [23] (panels c), the solar wind speed as measured by the Ulysses Solar Wind Plasma experiment (SWOOPS) [1] (panels d), the monthly sunspot number (hatched area) and the Ulysses heliographic latitude and heliocentric distance (panels e). The top graph ranges from 22 Oct 1992 to 30 Oct 1998, and the bottom graph from 30 Oct 1998 to 6 Jan 2005. During these two time intervals, Ulysses scanned the same heliolatitudes and helioradii but under different solar conditions. The first Ulysses orbit occurred during the decaying phase of the solar cycle 22 and rising phase of cycle 23; whereas the second orbit occurred during the maximum of solar cycle 23. Under solar minimum conditions (1st orbit), the low-latitude (<10◦ ) heliosphere was dominated by slow (∼400 km s−1 ) solar wind and the highlatitude heliosphere (>40◦ ) by fast (>700 km s−1 ). The mid-latitude (<30◦ ) heliosphere was dominated by the interaction between high-speed and lowspeed solar wind streams, producing CIRs able to accelerate electrons up to 300 keV and ions up to a few MeV/nucleon. The descent to southern heliolatitudes (years 1992-94) was characterized by regular increases of low-energy electron and ion intensities, at the solar rotation period (26 days sidereal), and observed up to S80◦ . Particles accelerated in the middle heliosphere at CIRs were are able to propagate to high heliolatitudes. The intensity of the CIR-events decreased as Ulysses moved to high latitudes. When Ulysses returned to low heliospheric latitudes (<40◦ ) during the fast-latitude scan (mid 1995) and later in 1996, the recurrent low-energy events reappeared. Intense SEP events were not observed until Nov 1997 coinciding with the rising phase of solar cycle 23 [10]. Panel c shows that very few SEP events (Oct 1992, Nov 1997, Apr-May 1998, Nov 1998) were observed during the first orbit at high energies. The background level of the 71-94 MeV proton channel was dominated by the increasing level of penetrating GCRs. This level started to decrease in Nov 1997 with the increase of the solar activity level. During solar maximum, solar wind was predominantly slow or mid-speed −1 (< ) at all heliolatitudes with a more frequent observation of CMEs ∼ 600 km s [16]. High-speed solar wind was observed only during the northern polar pass (late 2001). The intensity profiles showed numerous transient SEP events at both low (panels a,b) and high energies (panel c). Intense SEP events were observed even at the highest heliolatitudes [12]. Elevated low-energy intensity fluxes were observed throughout the second orbit. The complex pattern of the energetic particle profiles during the second Ulysses orbit was the result of the increasing level of solar activity and the dynamic evolution of the solar Heliospheric energetic particle variability 301 Fig. 3. Daily averages of (a) 40-65 keV electron, (b) 1.8-4.7 MeV ion, and (c) 71-94 MeV proton intensities, and (d) solar wind speed. Panel (e) shows the heliographic latitude (solid line), the helioradius (dashed line) of the Ulysses spacecraft, and the monthly sunspot number (hatched area). The top (bottom) graph covers the first (second) Ulysses orbit. Vertical dashed lines in the top graph are 26-days apart and indicate the solar rotation period. Vertical gray bars indicate the Ulysses polar passes at heliographic latitudes above 70◦ . 302 D. Lario corona and heliospheric structure. Small events associated with weak CIRs were also observed during the second orbit, although their observation was sporadic without showing any clear pattern [11]. 5 Summary Energetic particle observations during solar minimum and solar maximum are a faithful reflection of our changing Sun. During solar maximum, the Sun becomes the dominant source of energetic particles in the heliosphere, whereas during solar minimum the low-energy population mainly originates in the heliosphere by CIR events. The Ulysses mission has shown that during solar minimum CIR-associated energetic particles are easily transported from low to high latitudes, whereas at solar maximum SEPs are easily observed filling broad regions of the heliosphere. The complex structure of the heliosphere during solar maximum leads to a reduction of GCR fluxes and elevated fluxes of low-energy particles forming reservoirs in the inner heliosphere that gradually empty as the level of solar activity decreases. References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. S.J. Bame, D.J. McComas, B.L. Barraclough et al: A&A 92, 237 (1992) E.W. Cliver, A.G. Ling: Astrophys. J. 556, 432 (2001) A.C. Cummings, E.C. Stone: Space Sci. Rev. 83, 51 (1998) M.I. Desai, G.M. Mason, J.E. Mazur et al: Astrophys. J. 645, L81 (2006) J.H. Eraker, J.A. Simpson: Astrophys. J. 232, L131 (1979) P. Evenson: Space Sci. Rev. 83, 63 (1998) R.E. Gold, S.M. Krimigis, S.E. Hawkins et al: Space Sci. Rev. 86, 541 (1998) J.R. Jokipii, B. Thomas: Astrophys. J. 243, 1115 (1981) L.J. Lanzerotti, R.E. Gold, K.A. Anderson et al: A&A 92, 349 (1992) D. Lario, R.G. Marsden, T. Sanderson et al: J. Geophys. Res. 105, 18251 (2000) D. Lario, E.C. Roelof, R.J. Forsyth et al: Space Sci. Rev. 97, 249 (2001) D. Lario, E.C. Roelof, R.B. Decker et al: Adv. Space Res. 32, 579 (2003) D. Lario, E.C. Roelof, R.B. Decker et al: Ann. Geophys. 21, 1229 (2003) R.P. Lin: Solar Phys. 100, 537 (1985) G.M. Mason: Composition and energy spectra of ions accelerated in corotating interaction regions. In: Acceleration and Transport of Energetic Particles Observed in the Heliosphere: ACE 2000 Symposium, ed by R.A. Mewaldt et al. (AIP Conf. Proc. 528) pp 234–241 (2000) D.J. McComas, R. Goldstein, J.T. Gosling et al: Space Sci. Rev. 97, 99 (2001) F.B. McDonald: Space Sci. Rev. 83, 33 (1998) E.C. Roelof, R.E. Gold, G.M. Simnett et al: Geophys. Res. Lett. 19, 1243 (1992) T. Sanderson, V. Bothmer et al: Geophys. Res. Lett. 22, 3357 (1995) E.T. Sarris, S.M. Krimigis, T.P. Armstrong: J. Geophys. Res. 81, 2341 (1976) M.A. Shea, D.F. Smart: 27th Int. Cosmic Ray Conf. 8, 3401 (2001) G.M. Simnett, E.C. Roelof: Space Sci. Rev. 72, 303 (1995) J.A. Simpson, J.D. Anglin, A. Balogh et al: A&A 92, 365 (1992) K.D.C. Simunac, T.P. Armstrong: J. Geophys. Res. 109, A10101 (2004) Two years of Saturn’s exploration by the Cassini spacecraft: atmospheric studies A. Sánchez-Lavega, R. Hueso and S. Pérez-Hoyos Grupo Ciencias Planetarias, Dpto. Fı́sica Aplicada I, Universidad del Paı́s Vasco, agustin.sanchez@ehu.es Summary. This work reviews the exploration of Saturn’s atmosphere performed by the Cassini spacecraft in 2004-05. 1 Introduction Jupiter and Saturn are giant gaseous planets similar in their size, rotation rate and averaged physical and chemical properties. Both have a rich meteorology at the upper cloud level (pressures from a few mbar to about 3 bar) as observed in the visible and thermal infrared wavelength ranges (wavelengths 200 nm to 5 µm), and in the stratosphere and upper troposphere as sensed in the thermal infrared (wavelengths 7.8 - 50 µm) and using radio occultation methods. Atmospheric observations by Cassini spacecraft cover all this spectral range being performed with a variety of instruments. The ”Imaging Science Subsystem” (ISS) is a two camera system formed by a narrow angle telescope (2 m focal, 0.35◦ FOV) and a wide angle telescope (0.2 m focal, 3.5◦ FOV), both using CCD detectors (1024 square array of pixels), and a large set of filters (wide and narrow) ranging from UV (264 nm) to near IR (1012 nm)[21]. The ”Visual Infrared Mapping Spectrometer” (VIMS) is a multi-channel spectrometer which simultaneously acquires 352 bandpasses ranging from 0.35 to 5.1 µm, with two separated channels: visual (CCD detector) and near-IR (single element detector) [4]. The ”Composite Infrared Spectrometer” (CIRS) consists of two Fourier-transform spectrometers covering the wavelength range 7 µm - 1 mm at a high resolving power [7]. The ”Radio and Wave Plasma Science” (RPWS) experiment is designed to measure the electromagnetic radio emissions from 1 Hz to 16 MHz able to detect radio bursts from lightning activity and to extract the vertical temperature profiles from radio occultation investigations [13]. Cassini atmospheric studies focus on several topics. A major unresolved puzzle in these atmospheres is the nature of the system of zonal jet winds alternating in the East-West direction with latitude. A particular mystery is 304 A. Sánchez-Lavega, R. Hueso and S. Pérez-Hoyos the mechanism giving raise to the broad and intense equatorial jets on both planets [16, 38]. Globally Jupiter’s jets are stable in time in their peak latitude location and intensity within about ten per cent, but little is known about the temporal stability of Saturn’s jet system. Another important subject of research is the nature of the meteorological phenomena that are observed in the stratosphere and upper troposphere (vortices, waves and convective storms). The vertical and latitudinal structure mechanisms that form the hazes and clouds, the origin of the chromophore agent that gives the color to the clouds, and the temperature maps and the chemical distribution of non homogeneous compounds are other scientific objectives in the atmospheric studies of Saturn. 2 Saturn meteorology Saturn’s upper troposphere is expected to contain three main cloud layers according to thermochemical modelling [40] and radio observations [3, 12]. They are composed of ammonia (pressure level P ∼ 1 - 1.4 bar), ammonium hydrosulfide (NH4 SH, P ∼ 2-4 bar) and water (P ∼ 8-10 bar). Above the ammonia cloud, thick hazes extend up to few mbars. Radiative transfer modelling in the visible wavelengths allows to retrieve the upper cloud structure [37, 17, 20]. Most atmospheric features seen at visual wavelengths locate above the 1 bar level, and somewhere between 2-4 bar those sensed in the infrared window at ∼ 5µm. 2.1 Temperature and composition The temperature and chemical composition of the lower stratosphere and upper troposphere are being analyzed by the CIRS instrument and to some extent by the RPWS in the radio occultation mode. On the one hand, the meridional gradient of the stratospheric temperatures so far measured from 0.1 mbar to 0.5 bar imply a strong decay of the equatorial winds with altitude and a warm south polar stratosphere, as expected from seasonal insolation changes [8]. The temperature data also suggest vertical motions (upwelling in bands and downwelling in belts), with intense upwelling at equator. On the other hand, the spatial variations of PH3 and para-H2 indicate overturning motions in the troposphere (Hadley-like), and the NH3 distribution suggests different spatial levels of condensation [9]. The elemental abundances measurements indicate that C and P are both enriched but N is depleted with respect to protosolar values [9]. The C/H ratio is seven times solar, twice that of Jupiter. 2.2 Dynamical phenomena Vortices Saturn closed vortices have an oval shape and low albedo contrast as first shown by Voyager 1 and 2, and most have anticyclonic vorticity [34, 35, 36]. Cassini spacecraft atmospheric studies 305 Fig. 1. Examples of Saturn’s meteorological phenomena captured with Cassini ISS: Vortices and bright clouds (left: PIA 06507 on Sept. 19, 2004, filter CB2; lower right: PIA 07564 on July 6, 2005, filter CB2), and a convective storm with associated burst radio emission (lower right: PIA 07789 on January 27, 2006, filter CB2). The Cassini ISS images show also vortices in different latitudes of the northern hemisphere (fig. 1) [22, 39], some of them merging after their close encounter. They seem to be ubiquitous since they were detected as spots in HST images [23]. Much work remains to be done on the measurement of their properties (wind field, temperature and chemistry). Convective storms Bright, irregular in shape, mid-scale clouds that evolve rapidly, were also common in the Voyager era at mid-latitudes and in the equatorial area, referred here as ”plumes” [34, 35, 36, 25]. They are usually interpreted as the result of wet convection on the ammonia or water clouds [26]. The most dramatic convective events are the rare ”Great White Spots” (GWS) that are largescale phenomena disturbing completely the zone where they emerge. The most recent events occurred in 1990 [27] and 1994-95 [28] at Equatorial latitudes, between the Voyagers and Cassini visits. The ISS images show a continuous activity of several smaller storms at mid northern latitudes (fig. 1), and radio emissions episodes (Saturn electrostatic discharges, SED) associated to lightning in these storms have been detected by the RPWS [22, 14, 6] Waves at cloud level During the Voyager encounters two distinct waves covering the full circumference (latitude circle) were detected at cloud level: the so-called ”ribbon” 306 A. Sánchez-Lavega, R. Hueso and S. Pérez-Hoyos Fig. 2. Left: Saturn’s North Polar Region showing the hexagonal wave at the time of Voyager 2 (1979) [10]. Right: Saturn’s South Pole in a map composed Cassini ISS CB2 images [24]. at mid-northern latitudes [36] and the ”hexagon” close to the north pole [10] (fig. 2), re-observed later using the HST and ground-based facilities [29, 30]. Therefore they are long-lived waves and potential targets for Cassini ISS. Their nature is not yet well established although it has been proposed that they are Rossby waves evolving in a sheared flow [1, 11]. Other less prominent waves were seen in a reanalysis of the Voyagers southern hemisphere images [25], and some waves, extending along short latitude sectors, are evident in Cassini images. The first released images and results using VIMS are strongly promising in retrieving the meteorological phenomena and winds in the mid-altitude cloud (NH4 SH), as never seen before (fig. 3) [2, 18]. 3 Winds and circulation Saturn has a broad and intense zonal jet extending between latitudes ±40◦ blowing eastward with maximum speed at cloud level of 475 ms−1 as measured during the Voyagers encounters in 1980-81 [15, 25]. The jet showed much lower speeds during 1996-2004 [31, 23] and vertical wind shears were proposed to explain the velocity differences [22]. However, a new analysis of Saturn’s cloud vertical structure during the Voyager encounters suggested that real changes have occurred in the jet at cloud level [19]. Distinguishing between dynamical variability (temporal changes) and permanent vertical wind shears or the existence of both, is a fundamental step to understand the mechanisms and energy sources, external (solar) or internal (deep convection), that intervene in creating this strong jet. It must be mentioned that the velocities are measured Cassini spacecraft atmospheric studies 307 Fig. 3. Images of the deep cloud layer observed at ∼ 5 µm with VIMS. The top image was acquired on Feb. 17, 2005, the second image on March 8, 2005, and the third on July 12, 2005. The lower image (PIA 01941) was acquired on April 27, 2006. relative to the System III reference frame determined by the radio rotation period (assumed to be tied to the deep interior) as measured by the Voyagers [5]. However Ulysses and Cassini measurements of radio emission recurrence and, in this last case, of the magnetic field rotation, show variability in this period by at least several minutes, so it remains to be determined what is the Saturn rotation period [32]. 308 A. Sánchez-Lavega, R. Hueso and S. Pérez-Hoyos Fig. 4. Saturn’s zonal wind profile from two missions: grey line and dots from Voyager 1 and 2 in 1980-81 [25]; grey line with triangles and black line with dots from Cassini ISS images in 2004-05 [24, 33]. Note the profile differences at equator. The ISS camera is taking images of the planet regularly since 2003, but in particular since April 2004, before orbit injection. The images allow to track the atmospheric features motions [22, 39, 24, 33] (see fig. 2) and when properly calibrated in intensity, to infer the upper cloud and haze vertical structure using radiative transfer models [24, 33]. The first published study of motions in Saturn’s atmosphere using ISS images [22] showed two distinct patterns of zonal wind velocities at southern Equatorial latitudes based on the tracking of about 40 features in two filters: (a) The methane band MT2 (727 nm) sensed higher altitudes and detected low speeds; (b) The CB2 continuum filter (750 nm) sensed lower altitudes and detected higher speeds. These results are in agreement with the vertical wind shears derived above the clouds from temperature measurements and application of the thermal wind relationship [8]. We have extended this study increasing the temporal base, the wavelength Cassini spacecraft atmospheric studies 309 coverage and the number of targets. Our results show that between latitudes 5◦ N and 12◦ S the winds increase their velocity with depth from 265 ms−1 at the 50 mbar pressure level to 365 ms−1 at 700 mbar. These values are below the high wind speeds of 475 ms−1 measured at these latitudes during the Voyager era in 1980-81, indicating that the equatorial jet has suffered a significant intensity change between that period and 1996-2005 or that the tracers of the flow used in the Voyager images were rooted at deeper levels than those in Cassini images (fig. 4). Other works studied the motions at non Equatorial latitudes [39], mainly the temperate latitudes, where the winds showed good agreement with Voyager 1 and 2 (1980-81) and our HST data (1996-2004). One important result we have obtained analyzing Cassini ISS images was the discovery of a very strong vortex encircled by an intense jet with maximum speeds of 160 ms−1 at the south pole ([24]; fig. 2 and 4). 4 Future prospects Ongoing analysis and future additional images and measurements with the battery of atmospheric Cassini instruments should give precise information on the 3D temperature structure, aerosol properties and their vertical distribution, and chemical abundances and distributions of the different species (including aerosols) in both hemispheres, as well as their temporal variability. Cloud motions from the near infrared imaging with VIMS in the 4-5 micron window, with deeper penetration than in the visible, and the determination of the precise altitude location of the tracers at these wavelengths will complete the full 3D structure of the winds in Saturn’s upper troposphere. Acknowledgments: This work has been supported by the Plan Nacional de Astronoma y Astrofsica (MEC) AYA 2003-03216 and Grupos UPV 15946/2004. R. Hueso was supported by the Ramón y Cajal Program (MEC). This research made use of the public Cassini images available at NASA Planetary Data System. References 1. 2. 3. 4. 5. 6. 7. 8. 9. M. Allison, D. A. Godfrey, R. F. Beebe: Science 247, 1061 (1990) K. Baines et al.: Earth Moon Plan. 96, 119 (2005) F.H. Briggs, P.D. Sackett: Icarus 80, 77 (1989) Brown, R. A. et al.: Space Sci. Rev., 115, 111 (2004) M.D. Desch, M.L. Kaiser: Geophys. Res. Lett., 8, 253 (1981) Fischer et al.: Icarus 183, 135 (2006) F.M. Flasar et al.: Space Sci. Rev. 115, 169 (2004) F.M. Flasar et al.: Science 307, 1247 (2005) L.N. Fletcher, P.G.J. Irwin, N. Teanby, G.S. Orton, P.D. Parrish, R. de Kok, S.B. Calcutt, C. Howett, P.L. Read, F.W. Taylor and the Cassini/CIRS Team: Bull. Amer. Astron. Soc. 38, 501 (2006) 310 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. A. Sánchez-Lavega, R. Hueso and S. Pérez-Hoyos D.A. Godfrey: Icarus 76, 335 (1988) D.A. Godfrey, V. Moore: Icarus 68, 313 (1986) A.W. Grossman, D.O. Muhleman, G.L. Berge: Science 245, 1211 (1989) D.A. Gurnett et al.: Space Sci. Rev. 114, 395 (2004) D.A. Gurnett et al.: Science 307, 1255 (2005) A.P. Ingersoll, R.F. Beebe, B.J. Conrath, G.E. Hunt: Structure and dynamics of Saturn’s atmosphere. In: Saturn, (University of Arizona Press, Tucson, Arizona 1984) pp 195–238 A.P. Ingersoll, T.E. Dowling, P.J. Gierasch, G.S. Orton, P.L. Read, A. SánchezLavega, A.P. Showman, A.A. Simon-Miller, A.R. Vasavada: Dynamics of Jupiter’s Atmosphere. In Jupiter: The Planet, Satellites & Magnetosphere (Cambridge University Press, Cambridge, MA. 2004) pp 105–128 E. Karkoschka, M.G. Tomasko: Icarus 106, 428 (1993) T.W. Momary et al.: Bull. Amer. Astron. Soc., 38, 499 (2006) S. Pérez-Hoyos, A. Sánchez-Lavega: Icarus 180, 161 (2006) S. Pérez-Hoyos, A. Sánchez-Lavega, R.G. French, J.F. Rojas: Icarus 176, 155 (2005) C.C. Porco et al.: Space Sci. Rev. 115, 363 (2004) C.C. Porco et al.: Science 307, 1243 (2005) A. Sánchez-Lavega, R. Hueso, S. Pérez-Hoyos, J.F. Rojas, R.G. French: Icarus 170, 519 (2004) A. Sánchez-Lavega, R. Hueso, S. Pérez-Hoyos, J.F. Rojas: Icarus 184, 524 (2006) A. Sánchez-Lavega, J.F. Rojas, P.V. Sada: Icarus 147, 405 (2000) R. Hueso, A. Sánchez-Lavega: Icarus 172, 255 (2004) A. Sánchez-Lavega, F. Colas, J. Lecacheux, P. Laques, I. Miyazaki, D. Parker: Nature 353, 397 (1991) A. Sánchez-Lavega, J. Lecacheux, J.M. Gomez, F. Colas, P. Laques, K. Noll, D. Gilmore, I. Miyazaki, D. and Parker: Science 271, 631 (1996) A. Sánchez Lavega: Icarus 158, 272 (2002). A. Sánchez-Lavega A., S. Pérez-Hoyos, J. R. Acarreta, R. French: Icarus 160, 216 (2002) A. Sánchez-Lavega, S. Pérez-Hoyos, J.F. Rojas, R. Hueso, R.G. French: Nature 423, 623(2003) A. Sánchez-Lavega: Science 307, 1223 (2005) A. Sánchez-Lavega, R. Hueso, S. Pérez-Hoyos: Icarus, in press (2007) B.A. Smith et al.: Science 212, 163 (1981) B.A. Smith et al.: Science 215, 504 (1982) L.A. Sromovsky, H.E. Revercomb, R.J. Krauss, V.E. Suomi: J. Geophys. Res. 88, 8650 (1983) M.G. Tomasko, R.A. R.A. West, G.S. Orton, V.G. Tejfel: Cloud and aerosols in Saturn’s atmosphere. In: Venus, (University of Arizona Press, Tucson, Arizona 1984) pp 150–194 A.R. Vasavada, A.P. Showman: Rep. Progr. Phys. 68, 1935 (2005) A.R. Vasavada, S.M. Hrst, M.R. Kennedy, A.P. Ingersoll, C.C. Porco, A.D. Del Genio, R.A. West: J. Geophys. Res. 111 (2006) S.J. Weidenschilling, J.S. Lewis: Icarus 20, 465 (1974) A New Way for Exploring Solar and Stellar Magnetic Fields J. Trujillo Bueno1,2 1 2 Instituto de Astrofı́sica de Canarias, 38205 La Laguna, Tenerife, Spain, jtb@iac.es Consejo Superior de Investigaciones Cientı́ficas, Spain Summary. This paper highlights the diagnostic potential of some unfamiliar physical mechanisms that produce polarization in spectral lines, emphasizing how their exploitation could help us to map the “hidden” magnetic fields of the extended solar atmosphere. 1 Introduction Probably, the most interesting aspect of spectropolarimetry is that it allows us to diagnose magnetic fields in astrophysics. Cosmical magnetic fields leave their “fingerprints” in the state of polarization of the electromagnetic radiation that we collect with our increasingly large telescopes. This occurs through a variety of rather unfamiliar physical mechanisms, not only via the Zeeman effect. In particular, in stellar atmospheres there is a more fundamental mechanism producing polarization in spectral lines. There, where light escapes through the stellar “surface”, the atoms and molecules are illuminated by an anisotropic radiation field. The ensuing radiation pumping produces population imbalances among the magnetic substates of the energy levels (that is, atomic level polarization). As a result, the emission process can generate polarization in spectral lines without the need for a magnetic field. This is known as scattering line polarization (e.g., Stenflo 1994). However, light polarization components will also be selectively absorbed when the lower level of the transition is polarized (Trujillo Bueno & Landi Degl’Innocenti 1997; Manso Sainz & Trujillo Bueno 2003). Thus, the medium becomes dichroic simply because the light itself is escaping from it. In summary, the mere presence of atomic level polarization can generate spectral line polarization. The interesting point is that a magnetic field modifies the atomic level polarization via the Hanle effect, where the magnetic field B (in G) that produces a significant change is (e.g., Trujillo Bueno 2003a) B≈1.137×10−7 /(tlife gJ ) (1) 312 J. Trujillo Bueno (tlife and gJ being the lifetime, in seconds, of the J-level under consideration and its Landé factor, respectively). Since the lifetimes of the upper levels (Ju ) of the transitions of interest are usually much smaller than those of the lower levels (Jl ), it is clear that diagnostic techniques based on the lower-level Hanle effect are sensitive to much weaker fields than those based on the upper-level Hanle effect. 2 Scattering Physics and the Hanle Effect The spectral line polarization induced by the Zeeman effect is caused by the wavelength shifts between the π (∆M = Mu − Ml = 0) and σb,r (∆M = ±1) transitions. Therefore, the advantage of the Zeeman effect as a diagnostic tool is that the mere detection of polarization implies the presence of a magnetic field. The bad news are the following: • It is of limited practical interest for the determination of magnetic fields in hot coronal plasmas because the Zeeman polarization scales with the ratio between the Zeeman splitting and the Doppler-broadened line width. • The Zeeman effect is blind to magnetic fields that are tangled on scales too small to be resolved. Concerning the Hanle effect, these are the good news: • As shown by Eq. (1), the Hanle effect is sensitive to magnetic fields for which the Zeeman splitting in frequency units is comparable to the inverse lifetime of the upper (or lower) level of the spectral line used, regardless of how large the line width due to Doppler broadening is. It is therefore sensitive to weaker magnetic fields than the Zeeman effect: from at least 1 milligauss to hundreds of gauss. • It is sensitive to the presence of hidden, mixed-polarity fields at subresolution scales. • Contrary to a widespread belief, the diagnostic use of the Hanle effect is not limited to a narrow solar limb zone. In particular, in forward scattering at disk center, the Hanle effect can create linear polarization, when in the presence of inclined magnetic fields. In order to clarify the last two advantages let us consider scattering processes in a Jl = 0 → Ju = 1 line transition for the following two geometries: 90◦ scattering and forward scattering (see also Landi Degl’Innocenti & Landolfi 2004; and note that J is the total angular momentum of the atomic or molecular level under consideration). 2.1 The Hanle effect in 90◦ scattering The left panels of Fig. 1 (hereafter Fig. 1a) illustrate the 90◦ scattering case, in the absence and in the presence of a magnetic field. For this geometry A New Way for Exploring Solar and Stellar Magnetic Fields 313 Fig. 1. Left panels: the 90◦ scattering case in the absence (top panel) and in the presence (bottom panels) of a deterministic magnetic field. Right panels: the forward scattering case, in the absence (top panel) and in the presence (bottom panel) of a deterministic magnetic field. the largest polarization amplitude occurs for the zero field reference case, with the direction of the linear polarization as indicated in the top panel (i.e, perpendicular to the scattering plane). The two lower panels illustrate what happens when the scattering processes take place in the presence of a magnetic field pointing (a) towards the observer (left panel) or (b) away from him/her (right panel). In both situations the polarization amplitude is reduced with respect to the previously discussed unmagnetized case. Moreover, the direction of the linear polarization is rotated with respect to the zero field case. Typically, this rotation is counterclockwise for case (a), but clockwise for case (b). Therefore, when opposite magnetic polarities coexist within the spatio-temporal resolution element of the observation the direction of the linear polarization is the same as in the top panel of Fig. 1a, simply because the rotation effect cancels out. However, the polarization amplitude is indeed reduced with respect to the zero field reference case, which provides an “observable” that can be used for obtaining empirical information on hidden, mixed polarity fields at subresolution scales in the solar atmosphere (Stenflo 1994). Recently, three-dimensional radiative transfer modeling of the scattering polarization observed in atomic and molecular lines has shown that in the quiet regions of the solar photosphere there is a vast amount of “hidden” magnetic energy and unsigned magnetic flux (with hBi ∼ 100 G). This hidden magnetic energy, localized in the (intergranular) downflowing plasma, is carried mainly by tangled fields at sub-resolution scales with strengths between the equipartition field values and ∼ 1 kG, and is more than sufficient to compensate the radiative energy losses of the solar outer atmosphere (see Trujillo Bueno et al. 2004; 2006). 314 J. Trujillo Bueno 2.2 The Hanle effect in forward scattering The right panels of Fig. 1 (hereafter Fig. 1b) illustrate the case of forward scattering, in the absence (top panel) and in the presence (bottom panel) of a magnetic field. In this geometry we have zero polarization for the unmagnetized reference case, while the largest linear polarization is found for “sufficiently strong” fields (i.e., for a magnetic strength such that the ensuing Zeeman splitting is much larger than the level’s natural width). In other words, in the presence of an inclined magnetic field that breaks the symmetry of the scattering polarization problem at the solar disk center, forward scattering processes can produce measurable linear polarization signals in spectral lines (Trujillo Bueno et al. 2002). 3 Chromospheric magnetic mappers The above-mentioned presence of a significant amount of unresolved magnetic fields in the “quiet” solar photosphere might have several important consequences for the overlying solar atmosphere, such as ubiquity of reconnecting current sheets and heating processes (e.g., Priest 2006). Therefore, it is now even more important to carry out detailed empirical investigations on the magnetism of the “quiet” solar chromosphere via a clever exploitation of a variety of subtle physical mechanisms by means of which a magnetic field can create and destroy spectral line polarization (e.g., Trujillo Bueno 2003b). The following two subsections present two promising examples of diagnostic tools for the exploration of chromospheric magnetism. 3.1 The Hanle effect in the Ca ii IR triplet Figure 2 shows results of multilevel radiative transfer calculations aimed at investigating the magnetic sensitivity of the scattering polarization in the IR triplet of ionized calcium in a given thermal model of the solar atmosphere. Interestingly, while the emergent fractional linear polarization (Q/I) in the 8498 Å line (with Jl = Ju = 3/2) shows a strong sensitivity to inclined magnetic fields with strengths between 1 and 10 G, the emergent Q/I in the 8542 Å (Jl = 5/2 and Ju = 3/2) and 8662 Å (Jl = 3/2 and Ju = 1/2) lines is very sensitive to magnetic fields in the milligauss range (i.e., between 10−3 and 10−1 G). The reason for this very interesting behavior is that the Q/I in the 8498 Å line is dominated by the selective emission processes that result from the atomic polarization of the short-lived upper level, while the Q/I in the 8542 Å and 8662 Å lines is dominated by the selective absorption processes that result from the atomic polarization of the long-lived lower levels (Manso Sainz & Trujillo Bueno 2003). Interestingly, the calculated polarization amplitudes are also very sensitive to the assumed thermal model of the solar chromosphere (see Fig. 4 of Manso A New Way for Exploring Solar and Stellar Magnetic Fields 315 Fig. 2. The fractional linear polarization amplitude of the Ca ii IR triplet calculated for a line-of-sight at µ = cosθ = 0.1 (i.e., at about 5 arc seconds from the limb, since θ is the heliocentric angle) in a simple model of the solar atmosphere. Each curve corresponds to the indicated inclination (θB ) of the assumed random-azimuth magnetic field. From Trujillo Bueno & Manso Sainz (2002). Sainz & Trujillo Bueno 2001). In conclusion, simultaneous observations of the scattering polarization in the Ca ii IR triplet should be carried out in order to use them as a sensitive thermometer and magnetometer of the quiet regions of the solar chromosphere. 3.2 Forward scattering in the He i 10830 Å multiplet As mentioned in Section 2, in the presence of an inclined magnetic field forward scattering processes can produce linear polarization signatures in spectral lines. In this geometry, the polarization signal is created by the Hanle effect, a physical phenomenon that has been clearly demonstrated via spectropolarimetry of solar coronal filaments in the He i 10830 Å multiplet (Trujillo Bueno et al. 2002). The linear polarization detected in the “blue” line of this multiplet (i.e., that with Jl = 1 and Ju = 0) demonstrated the operation of the lower-level Hanle effect at the solar disk center, while that measured in the “red” blended components (with the same Jl = 1, but with Ju1 = 2 and Ju2 = 1) showed an additional contribution caused by the upper-level Hanle effect. Fig. 3 shows theoretical examples of the emergent fractional linear polarization in the lines of the He i 10830 Å multiplet, assuming a constant-property slab of helium atoms permeated by a horizontal magnetic field of 10 G (see also the information provided by Trujillo Bueno & Asensio Ramos 2007). As expected, the smaller the optical thickness of the assumed plasma structure the smaller the fractional polarization amplitude. In principle, the Tenerife Infrared Polarimeter built by the IAC allows the detection of very low amplitude polarization signals, such as those corresponding to the ∆τ = 0.1 case 316 J. Trujillo Bueno Fig. 3. The emergent fractional polarization at the solar disk center in the lines of the He i 10830 Å multiplet, assuming that a constant-property slab of helium atoms at a height of about 2000 Km is permeated by a horizontal magnetic field of 10 G. The various Q/I profiles correspond to the slab’s optical thickness indicated in the inset, calculated at the wavelength of the red blended component. The positive reference direction for the definition of the Stokes Q parameter is along the magnetic field vector. From Asensio Ramos & Trujillo Bueno (2007; in preparation). of Fig. 3. However, in order to be able to achieve this goal without having to sacrifice the spatial and/or temporal resolution we need a large aperture solar telescope. Certainly, a promising investigation would be to explore the topology of the chromospheric magnetic field via this novel diagnostic window. Of particular interest would be the detection of “magnetic canopies” in the quiet solar chromosphere or, alternatively, lack of a clear observational evidence for the presence of such large areas of diffuse, predominantly horizontal magnetic fields. 4 Coronal magnetic mappers The intensity profiles of the hydrogen lines of the Lyman series have been measured on the solar disk by several instruments on board rockets and space telescopes (e.g., by the SUMER spectrometer on SOHO), showing that such lines are in emission at all positions and times and that they originate in the upper chromosphere and transition region (e.g., Warren et al. 1998). As predicted by Eq. (1), the scattering polarization of the hydrogen lines of the Lyman series should be sensitive (via the Hanle effect) to the typical magnetic A New Way for Exploring Solar and Stellar Magnetic Fields 317 Fig. 4. Theoretical estimate of the linear polarization created by the Hanle effect in the Lyman α line as a result of forward scattering processes in the presence of a horizontal magnetic field in the solar transition region. Note that the Q/I amplitude increases with the strength of the magnetic field, up to the Hanle-effect saturation field intensity (∼ 200 G for the forward scattering case in the presence of a horizontal field). The calculations have been carried out assuming a slab of hydrogen atoms at a height of 3000 Km above the solar visible surface and neglecting the contribution of the center-to-limb variation of the Lyman α radiation on the anisotropy factor. From Trujillo Bueno et al. (2005). strengths expected for the solar outer atmosphere (chromosphere, transition region and corona).Therefore, it is of great interest to investigate the diagnostic potential of the Hanle effect in the hydrogen lines of the Lyman series for the forward scattering case. Figure 4 shows a theoretical example of the linear polarization signal created by the Hanle effect of a horizontal magnetic field in the Lyman α line. In conclusion, these forward scattering polarization signals offer a novel diagnostic tool for mapping the magnetic fields of the solar transition region and lower corona. Obviously, the required observations can be realized only from a UV/EUV spectropolarimeter on board of a space telescope. 5 Concluding remarks As we have seen, the observation and theoretical modeling of spectral line polarization produced by the presence of atomic level polarization offers a very attractive diagnostic window for the exploration of solar and stellar magnetism. Of particular interest are the new diagnostic opportunities we now have for mapping the “hidden” magnetic fields of the quiet solar photosphere, chromosphere and corona. To this end, it is first necessary to develop a large aperture ground-based solar telescope optimized for spectropolarimetric observations and to put a high-sensitivity UV/EUV polarimeter in a space tele- 318 J. Trujillo Bueno scope. It is also urgent because the Sun must be viewed, once again, as a Rosetta stone for astrophysics. Acknowledgements: Finantial support by the Spanish Ministerio de Educación y Ciencia through project AYA2004-05792 is gratefully acknowledged. References 1. Landi Degl’Innocenti, E., Landolfi, M.: Polarization in Spectral Lines, (Kluwer Academic Publishers, 2004) 2. Manso Sainz, R., Trujillo Bueno, J.: Modeling the Scattering Line Polarization of the Ca ii IR Triplet. In: Advanced Solar Polarimetry: Theory, Observation and Instrumentation, ed by M. Sigwarth (ASP Conf. Series Vol. 236, 2001) pp 213–220 3. Manso Sainz, R., Trujillo Bueno, J.: Phys. Rev. Letters 91, 111102-1 (2003) 4. Priest, E.: Our Magnetic Sun. In: The Many Scales of the Universe, ed by J. C. del Toro Iniesta et al. (Kluwer Academic Publishers, 2006) pp 197 – 211 5. Stenflo, J. O.: Solar Magnetic Fields, (Kluwer Academic Publishers, 1994) 6. Trujillo Bueno, J.: The Generation and Transfer of Polarized Radiation in Stellar Atmospheres. In: Stellar Atmosphere Modeling, ed by I. Hubeny, D. Mihalas and K. Werner (ASP Conf. Series Vol. 288, 2003a) pp 551–582 7. Trujillo Bueno, J.: New Diagnostic Windows on the Weak Magnetism of the Solar Atmosphere. In: Solar Polarization 3, ed by J. Trujillo Bueno, J. Sánchez Almeida (ASP Conf. Series Vol. 307, 2003b) pp 407–424 8. Trujillo Bueno, J., Landi Degl’Innocenti, E.,: ApJ Letters 482, L183 (1997) 9. Trujillo Bueno, J., Manso Sainz, R.,: Il Nuovo Cimento Vol. 25 C, No. 5-6, 783 (2002) 10. Trujillo Bueno, J., Landi Degl’Innocenti, E., Collados, M., Merenda, L., Manso Sainz, R.: Nature 415, 403 (2002) 11. Trujillo Bueno, J., Shchukina, N., & Asensio Ramos, A.: Nature, 430, 326 (2004) 12. Trujillo Bueno, J., Landi Degl’Innocenti, E., Casini, R., Martı́nez Pillet, V.: The Scientific Case for Spectropolarimetry from Space. In: 39th ESLAB Symposium on Trends in Space Science and Cosmic Vision 2020, ed by F. Favata et al. (ESA Publications Division, Vol. SP-588, 2005) pp 203–212. 13. Trujillo Bueno, J., Asensio Ramos, A., Shchukina, N.: The Hanle Effect in Atomic and Molecular Lines: A New Look at the Sun’s Hidden Magnetism. In: Solar Polarization 4, ed by R. Casini and B. W. Lites (ASP Conf. Series, Vol. in press, 2006) 14. Trujillo Bueno, J., Asensio Ramos, A.: ApJ 655, in press (2007) 15. Warren, H. P., Mariska, J. T., & Wilhelm, K.: ApJS, 119, 105 (1998) Session VII Observatories and instrumentation The MAGIC Telescopes (and beyond...) M. Martı́nez for the MAGIC Collaboration Institut de Fisica d’Altes Energies (IFAE), Edifici Cn. UAB, 08193 Bellaterra (Barcelona), Spain, martinez@ifae.es Summary. The MAGIC Telescope, located at the Roque de los Muchachos Observatory at the Canary Island of La Palma is at present the largest and most sensitive single telescope worlwide for the observations of very-high-energy cosmic gamma rays and incorporates pioneering technologies in many of its elements. In its design, construction, commissioning and scientific exploitation Spanish groups are playing a very important role. The technical aspects of the MAGIC construction and operation, its observational results and the status of the upgrade of the installation with a second, even more advanced, telescope are discussed. 1 The MAGIC telescope 1.1 Introduction The MAGIC telescope is a single mirror instrument designed to detect the Cherenkov light produced by the electromagnetic cascades initiated by incoming high energy gammas on the upper atmosphere. The high energy gammas (50 Gev-10 Tev) passing through high atmosphere start an electromagnetic shower from an altitude of 15 Km. The main idea is to use the atmosphere as a giant electromagnetic calorimeter to detect the energy and the incoming direction of cosmic gammas. The shower properties are reconstructed through the detection of the Cherenkov light emitted by superluminal electrons generated in the cascade. The light pool produced by the showers covers an area of typically 105 − 6 10 m2 , such large detection area is independent from the mirror size making the Cherenkov telescopes highly sensitive. The dish area indeed is important for the detection of the lower energy gammas. The MAGIC telescope is actually the largest telescope of this category of instruments and therefore its energy threshold is the lowest, 50 GeV, filling the gap of the covered spectrum between satellite experiments and other VHE gamma experiments. 322 M. Martı́nez for the MAGIC Collaboration The Telescope is located at an altitude of 2200 m in the area of the Observatorio del Roque de los Muchachos, in the Canary islands, covering mainly northern hemisphere sources, but allowing observations also of southern hemisphere candidates, as galactic center, with a higher energy threshold. The experiment was inaugurated in October 2003 and during 2004 started commissioning runs. In may the Cycle I of observations begun allowing the detection of known and new sources. 1.2 Detection principles The gamma primary high energy photons impinging in the upper atmosphere initiate a shower made exclusively by electron-positron couples and photons. The large parabolic mirror telescope, collects photons produced by superluminal charged particles and produces in the focus a typical image that develops in short time. The main source of background is generated by night sky light produced by stars and light scattered by the atmosphere. Fig. 1. Sketch of the principle of the Cherenkov technique through the formation of the image of an extended air shower in an imaging air Cherenkov telescope pixelized camera. The numbers in the figure correspond to a typical 1 TeV gamma-ray induced shower. The MAGIC Telescopes (and beyond...) 323 The other main source of background is given by showers initiated by different particles than gamma-rays. In particular high energy protons have a rate three orders of magnitude larger than the brightest gamma source and have an isotropic direction distribution. The main background for the lower energy gamma-rays is produced by large impact parameter muons that mimic gamma-like images reproducing also their temporal behavior. The missing of a clear signature for low energy (¡100 GeV) gamma images rises the analysis energy threshold, while the trigger threshold has been measured to be well below 50 GeV. 1.3 The telescope technology The MAGIC telescope has been developed with the target of lowering the energy threshold of the instrument. For this reason new technologies had been developed during the R&D period that has been proved crucial. In particular three main issues have been exploited to reach low energy threshold: high quantum efficiency photosensors [13], large lightweight space frame and mirrors [11], low integration time in both trigger [12] and data acquisition [10]. The main dish has a mean diameter of 17m for an equivalent reflecting surface of 234 m2 made by 936 square 0.5m × 0.5m mirrors. The mirrors are made by lightweight composite aluminum structure. The reflecting surface has been obtained milling a spherical shape on the front aluminum panel. The other goal of the experiment is to exploit observations on Gamma Ray Burst and the entire structure has to be repositioned in a short time from the alerts coming from satellites. For this reason the structure has been made with carbon fiber tubes leading to a maximum weight of 17 t. The supports are made of steel tubes and the final gross weight of the telescope is 65 t. Since the structure is not stiff enough to maintain accurately the reflector shape while changing zenith angle, the mirror panels are focused dynamically with an Active Mirror Control system. Each panel has a laser pointer that allows to measure the off pointing and a dynamical system, based on LookUp-Tables, continuously adjusting each panel for different zenith angles. The photosensors are 1 inch photomultipliers for the inner pixels and 1.5 inch tubes for the external low resolution pixels 2. The tubes have an hemispherical entrance window to enlarge the probability of double hit of the incoming photons and a wavelength shifter coating has been added on the surface that increases the mean quantum efficiency by 30% from the original one. The same coating allows sensitivity on the UV band as shown in fig 3. To preserve the signal integrity the transmission has been made with an analog optical system. A VCSEL laser diode transmits the signals over 175 m from the telescope camera to the counting house, where fast receiver boards convert back to electrical the information and prepare the signals for the trigger and data acquisition. 324 M. Martı́nez for the MAGIC Collaboration Fig. 2. Front view of the MAGIC camera during an access for maintenance at the telescope. The camera, all its auxiliary systems and its complex control software has been designed, build and is maintained by the IFAE team. Fig. 3. Quantum Efficiency measurements on 1 inch Electron Tube 9116A photomultiplier. Lower non coated PMT, upper coated PMT. To reduce at maximum the night sky background (NSB) noise, the trigger performs coincidences in windows of 6ns on selectable close compact images. The DAQ is made by 300 M Hz samplers that acquire a time window of 30ns to fully reconstruct the original signal. Since the sampling frequency is a little marginal compared with the expected width of the shower signals, a filter has been added to stretch up to 6 ns the peaks to allow a good measurement of the absolute arrival time of the light on the tube. A double gain system has been added to increase the dynamical range of the 8 bit FADC of the samplers. 2 Physics Results 2.1 Cycle 1 Observations From May 2005 the telescope started an observation campaign, named Cycle1, but many hours of observations on galactic and extragalactic sources where already available from commissioning phase. The data taken since the last months of 2004, where the telescope performed almost on it’s design expectations are included in the analysis presented here. During Cycle I where taken 750 hours of data from extragalactic sources and 500 hours from galactic sources, apart from Crab data used as standard candle to calibrate the instrument. In this phase we have monitored the performances of the telescope and measured the energy and angular resolution. The measurements show an energy resolution of 20% for higher energies increasing up to 35% for threshold energies, while the spatial resolution show a PSF of 0.10o for higher energies increasing to 0.13o for threshold energies. The Crab plerion has been observed during winter 2004 and 2005 revealing a gamma signal at the telescope of 20σ/h. The MAGIC Telescopes (and beyond...) 325 Fig. 4. Highlights of the MAGIC Cycle 1 observations. 2.2 Extragalactic sources The search for very high energy gamma-ray emission from Active Galactic Nuclei (AGNs) is one of the major goals for ground-based gamma-ray astronomy. The photon-photon interaction produces a cut-off for higher energies and a correlation between redshift and maximum energy of detected sources, known as Fazio-Stecker relation, is expected. The Cycle I strategy on this category of sources was to observe a sample of X-ray bright (F1keV > 2µJy) northern HBLs at moderate redshifts (z < 0.3). The sample of candidates was chosen based on predictions from models involving a Synchrotron-Self-Compton and hadronic origin of the gamma-rays In parallel with the observations of the AGNs with MAGIC, the sources were observed with the KVA 35 cm telescope, also located at La Palma. In addition to these joint campaigns, several AGNs are regularly observed as part of the Tuorla Observatory Blazar monitoring program with the Tuorla 1 m and the KVA 35 cm telescopes. During the entire observation period Mkn 421 [1] was found to be in a low to moderate high flux state ranging from 0.5 to 2 Crab units above 200 GeV. Significant variations of up to a factor of four overall and up to a factor two in between successive nights has been seen and a clear correlation between Xray (taken from the All-Sky-Monitor on-board the RXTE satellite) and VHE gamma-ray data is verified. Mkn 501 was observed for 24 nights (Table 1) by the MAGIC telescope. The source was found to be in a rather low flux state during most of the 326 M. Martı́nez for the MAGIC Collaboration Table 1. Summary of observation of Cycle I AGN survey. ZA is the zenith angle in degrees. EThr is the analysis threshold in GeV. T(hr) is the observation time in hours. Source Period Mkn 421 Mkr 501 1ES 2344 Mkn 180 1ES 1959 1ES 1218 PG 1553 PG 1553 Nov04-Apr05 Jun-Jul05 Sep-Dec05 Mar06 Sep-Oct06 Jan05 Apr-May05 Jan-Apr06 ZA ET hr T(hr) 9-55 10-31 23-34 39-44 36-46 2-13 12-30 20-30 150 150 160 200 180 140 140 140 25.6 29.7 27.6 11.1 6.0 .2 7.0 11.8 observations. The flux level above 200 GeV was 30%-50% of the Crab Nebula flux with a strong indication of an Inverse-Compton peak. On five nights, the source was found in a flaring state with the flux reaching up to 4 Crab units. Moreover, a rapid flare with a doubling time as short as 5 minutes or less was detected on the night of 10 July 2005. The rapid increase in the flux level was accompanied by a hardening of the differential spectrum. This is the first time that spectral hardening was detected on time scales of some 10 minutes. A detailed publication on the analysis and results on the observation of Mkn 501 is in preparation. 1ES 2344+514, reported a VHE gamma-ray emission by Whipple in 1996. The MAGIC observation of 1ES 2344+514 (Table 1) yielded a clear excess with a significance of 11.5σ . The differential energy spectrum can be fitted by a simple power law with a photon index of 2.96 ± 0.12. The AGN Mkn 180 [2] (1ES 1133+704) is a well-known HBL at a redshift of z = 0.045. The observation of Mkn 180 was triggered by a brightening of the source in the optical on March 23, 2006, detected by the KVA telescope. The alert was given as the core flux increased by 50% from its quiescent level value. Mkn 180 was observed by the MAGIC telescope in 2006 during 8 nights (Table 1). The signal of 165 excess events was found with a significance of 5.5σ. No evidence for flux variability was found. The observed integral flux above 200 GeV is 11% of the Crab Nebula flux. 1ES 1959+650 [3] was observed by MAGIC during the commissioning phase in 2004 (Table 1). The analysis of the data gave a detection of VHE gamma-rays with a significance of 8.2σ. MAGIC observed 1ES 1218+304 [4] in seven nights in January 2005 (Table 1). The observed excess of 560 events has a statistical significance of 6.4 standard deviations above 140 GeV. 1ES 1218+304 is the first source discovered by MAGIC. The gamma-ray lightcurve did not show signs of significant variability. The energy spectrum was fitted with a pure power law with a photon index of 3.0 ± 0.4. The MAGIC Telescopes (and beyond...) 327 PG 1553+113 [5] was observed with the MAGIC telescope in 2005 at about the same time as the H.E.S.S. observations took place. Motivated by a hint of a signal in the preliminary analysis of the MAGIC data, additional observations were performed in 2006 (Table 1). Combining the data from 2005 and 2006, a very clear signal was detected with a total significance of 8.8σ. 2.3 Galactic sources About 1/3 of the observation time (not counting that devoted to Crab nebula technical observations) was devoted to galactic objects. The observations covered both, candidates and well established VHE gamma-ray emitters, and included the following types of objects: supernova remnants (SNRs), pulsars, pulsar wind nebulae (PWN), microquasars (QSRs), the Galactic Center (GC), one unidentified TeV source and one cataclysmic variable. In this section we highlight the results obtained so far from such observations. We have confirmed the VHE gamma-ray emission from the SNRs HESS J1813-178 [6] and HESS J1834-087 (W41) [7]. We have also measured the VHE gamma-ray flux from the Galactic Center [8]. The possibility to indirectly detect dark matter through its annihilation into VHE gamma-rays has risen the interest to observe this region during the last years. Our observations have confirmed a point-like gamma-ray excess whose location is spatially consistent with Sgr A* as well as Sgr A East. The energy spectrum of the detected emission is well described by an unbroken power law of photon index α = −2.2 and intensity about 10% of that of the Crab nebula flux at 1 TeV. LS I +61 303 [9] was observed in the VHE regime with MAGIC during 54 hours (after standard quality selection, discarding bad weather data) between October 2005 and March 2006. The source shows an evident pulsed VHE emission in phase with orbital parameters. The maximum of the pulse is present between phase 0.4-0.7, while at periastron is absent. This is the first periodic signal seen from VHE emitters. 3 MAGIC-II The MAGIC telescope has demonstrated to work within the design specifications but the analysis on the low energy gamma-rays on the other hand has proven to be difficult due to the background. Presently we are upgrading the installation adding a mechanical clone of the MAGIC telescope, named MAGIC-II. The stereo images are proven to reduce the muon background and improve the energy and direction resolution. The two telescope design is foreseen to double the sensitivity. The new telescope will include many new technologies as: lighter 1 square meter mirrors, new photosensors and new faster samplers. The big improvement is expected by high quantum efficiency devices such as HPD’s and SiPM. A new 2GHz sampler, based on a custom 328 M. Martı́nez for the MAGIC Collaboration Fig. 5. Picture of the MAGIC Telescopes in summer 2006. The second one (in the front) will be instrumented during 2007. chip named ”Domino”, is foreseen to reduce the NSB contribution on the signal. The heat dissipation required by standard FADC is causing troubles and the new design will reduce it by 2 orders of magnitude. The MAGIC-II telescope 5 is expected to start it’s commissioning phase late 2007 and join MAGIC in the data taking in 2008. 4 And beyond.. The success of the new generation Cherenkov Telescopes such as H.E.S.S. and MAGIC in opening the VHE gamma-ray window proves that it is the right technology to progress even further in the exploration of the VHE universe. To follow the same progression experimented in X-ray astronomy or gamma-ray astronomy, a big installation of Cherenkov telescopes is being planned. The CTA (Cherenkov Telescope Array) should be an European-lead large array of Cherenkov Telescopes, distributed among two sites (one in the north and one in the south) to allow full sky coverage and aiming to increase the sensitivity by more than one order of magnitude and extend the sensitivity to about four decades in energy, with a large overlap with the energy range observable by satellites 6. The CTA is being design as an semi-open facility for the astrophysics community and should start construction already by 2010 allowing a the VHE gamma-ray domain to reach full maturity in the coming decade as has happened with other high energy astronomies during the last decades. Acknowledgements: I’m indebted to Dr. N. Turini for his help with this manuscript. The MAGIC Telescopes (and beyond...) 329 Fig. 6. The sensitivity aimed for the CTA installation compared with the sensitivity already attained by new-generation installations such as H.E.S.S. and MAGIC. References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. J. Albert et al,. submitted to ApJ 2006 J. Albert et al,.submitted to ApJ Letters in June 2006 J. Albert et al., ApJ 639 (2006) 761-765 J. Albert et al., ApJ Letters 642, L119 (2006) J. Albert et al,.submitted to ApJ Letters in May 2006 J. Albert et al., ApJ Letters 637, L41 (2006) J. Albert et al., ApJ Letters 643, L53 (2006) J. Albert et al., ApJ Letters 638, L101 (2006) J. Albert et al., Science 312, 1771 (2006) H. Bartko et al., astro-ph/0506459 C. Bigongiari et al., Nucl.Instrum.Meth. A 518 (2004) 193 M. Meucci, R. Paoletti et al., NIM A (2004), Issue 1-2, 554 D. Paneque et al., Nucl.Instrum.Meth. A 518 (2004) 619 Present and Future of Astronomy at the Observatorio del Teide A. Oscoz Instituto de Astrofı́sica de Canarias, E-38200, La Laguna, Tenerife, Spain, aoscoz@iac.es Summary. The Observatorio del Teide and the Observatorio del Roque de los Muchachos together contain the largest European group of facilities dedicated to astronomical observations. For many years Spain’s contribution to the Observatorio del Teide was limited to the use the foreign facilities or telescopes bequeathed by their previous owners. This situation has changed and nowadays numerous instruments, experiments and telescopes exist in which the Spanish participation is remarkable or even total. New facilities have been or are soon to be installed at the Observatory, and a summary of the present situation is given in this paper. 1 Introduction The Spanish and international astronomical communities have access to a series of facilities that allow them to make high level scientific research. The Observatorio del Teide (OT), on Tenerife, operative since 1964, houses one the most important European collections of facilities. Although the OT is a Spanish observatory, until a few years ago most of its facilities had been installed by other countries, with only a residual presence of the Spanish astronomy. However, this situation has gradually changed over time and is now quite different. The structure of the changes at the OT can be separated into several categories. Firstly, it is convenient to distinguish between the Instruments, Telescopes and Experiments (ITEs), and there are new elements in all three categories: telescopes, with GREGOR and STELLA, five state-of-the-art instruments for three of the telescopes and experiments, such as STARE, PASS and EAST. The other new points are related to a restructuring of the support staff of the Spanish facilities, funding for the access by the observers to the telescopes and services to improve the assisance the observers must receive. A brief explanation of all this news is given in the next sections. 332 A. Oscoz Fig. 1. The Observatorio del Teide (OT) and the situation of the various telescopes mentioned in this paper. 2 Telescopes There are two new facilities at the OT that will come into operation in 2007: GREGOR, a solar telescope, and STELLA, a combination of two nocturnal telescopes. 2.1 GREGOR GREGOR is a new solar telescope with a 1.5 m aperture presently being assembled at the OT. The project is being undertaken by a predominantly German consortium (the Kiepenheuer Institut für Sonnenphysik (KIS), the Astrophysikalisches Institut Potsdam (AIP) and the University of Göttingen) and other international Partners. Until the huge ATST comes into operation, GREGOR will be the largest telescope for solar studies. As in other facilities, 20% of its time goes for the Spanish astronomical community. It is designed for high precision measurements of the magnetic field and gas motion in the solar photosphere and chromosphere with a resolution of 70 km on the Sun, corresponding to an angle of 0.1 arcsec, and for high resolution stellar spectroscopy. With a 300 arcsec of field of view and a wavelength range from 0.39 Present and Future of Astronomy at the Observatorio del Teide 333 nm to about 2.5 nm, it is also designed for night-time observations. It will be equipped with a Fabry–Perot spectrometer and several slit spectrographs in the optical, UV and IR ranges. 2.2 STELLA STELLA (STELLar Activity) is a project for observing and monitoring activity tracers on cool stars with two 1.2 m robotic telescopes, STELLA-I and STELLA-II, which operate in fully unattended mode. STELLA-I, f/8 with a FOV of 24 arcminutes, fibrefeeds an echelle spectrograph and hosts an optical CCD imager and photometer. On the other hand, STELLA-II, f/10 with a FOV of 2 arcminutes, will also host an optical CCD imager and photometer, together with an AO testbed. STELLA is run by the AIP in collaboration with the Instituto de Astrofı́sica de Canarias (IAC), and once again 20% of the observing time is awarded to Spanish observers. 3 Instruments Besides group of new telescopes, a new generation of state-of-the-art instruments, both solar and nocturnal, will shortly become available at the OT. 3.1 The IAC80 telescope: CCD and TCP The IAC80 was completely designed and built by the IAC and was the first of this class developed in Spain. It has an equatorial German mount, with an effective focal ratio of f/11.3 and an 82 cm primary mirror. Since 2005, the IAC80’s common-user instrumentation has been a 2k×2k E2V back-illuminated thinned CCD with 13.5 µm pixels, equivalent to 0.33 arcsecond per pixel in the sky. The control of the CCD goes through a userfriendly interface developed at the IAC. Another frequently mounted instrument at the IAC80 is the TCP (Tromsoe CCD Photometer) camera, a portable instrument optimized for fast readout photometry based on CCD technology. The TCP was built by the Physics Department of Tromsoe University (Norway) together with the CUO (Copenhagen University Observatory). The TCP uses a back-illuminated SiTe 1k×1k chip with a pixel size of 24µm. It offers the possibility of on-line data reduction. The telescope would be complete and offer a total service to the astronomer needing an installation on which to observe during long periods of time if it had a spectrograph. The idea is to incorporate a spectrograph for the IAC80 during 2007, together with improvements in the acquisition and guiding system. 334 A. Oscoz 3.2 The Carlos Sánchez telescope (TCS): FIN and FastCam The TCS, dedicated to nocturnal observations in the infrared, has a primary mirror of diameter 1.52 m. Operative since 1972, originally it was designed as an infrared flux collector, being one of the first thin mirror telescopes. The TCS has been owned by the IAC since 1982. Its common-user instrumentation includes two new state-of-the-art instruments, besides the infrared camera CAIN. FIN is a fast and precise infrared photometer with an InSb detector for observations of pointlike objects in the range from 1 to 5 microns. FIN was installed in 2004, but two problems in the instrument made it impossible to get useful data until they have been solved. With a user-friendly interface, 12 filters and on-line data reduction, FIN is one of the few infrared photometers in existence today. A forthcoming, and probably the most important, new instrument will be FastCam. This project, born in March 2006 in collaboration with the Universidad Politécnica de Cartagena, consists of a prototype L3CCD camera with very fast readout (several hundreds of images per second), which will allow us to obtain a seeing close to the diffraction limit of the telescope. With this, both spatial and temporal resolution can be achieved, crucial for many research projects. Moreover, the package now being developed for this instrument allows it to select, recentre and combine the best images, so that the final results are really impressive. The first tests offered spectacular images, even under bad atmospheric conditions, and gave a seeing of 000 .16 in the Johnson–Bessell I band, something never before seen at the OT. This instrument will be exported in the near future to other telescopes of greater size. Fig. 2. An image of Z Bootis obtained with FastCam: two stars separated by only 0.5 arcsecond. Present and Future of Astronomy at the Observatorio del Teide 335 3.3 VTT: TIP and TIP-II The Vakuum Turm Teleskop (VTT) is a classical solar telescope: two coelostats at the top feed the sunlight into the telescope which spans 10 floors. The primary mirror has a diameter of 70 cm and a focal length of 46 m. The VTT is equipped with several focal plane instruments for the observation of objects and physical processes on the solar surface. One of these instruments is the Tenerife Infrared Polarimeter, developed at the IAC. TIP, a common-user instrumentation on the VTT since 1999, is integrated into the VTT spectrograph and performs spectropolarimetry in the near infrared. All wavelengths up to almost 2 microns can be recorded. Its spatial and spectral sampling are ∼000 .35 and 30 mA, respectively. Success in the use of TIP made it possible to built a new polarimeter, TIP-II, with a better sensor (TCM-8600 1024×1024 instead of NICMOS 3 256×256) and better spatial and spectral sampling: ∼000 .18 and 18 mA. TIP-II has been installed at the VTT since 2005, although its final destination will be the GREGOR. Once GREGOR starts its operations, the observers will be able to use TIP on the VTT and TIP-II on the GREGOR simultaneously . Fig. 3. Comparison the quality of TIP-I vs. TIP-II. 4 Small Experiments The Solar Laboratory is a peculiar telescopic installation at the OT owned by the IAC that contains a total of six instruments operating continuously (on a daily basis and in some cases for more than 25 years) both day and night 336 A. Oscoz and with a unique and precise scientific observing programme. The owners of the instruments are various but IAC researchers are responsible for their operation and actively participate in their scientific exploitation through the various established consortia. The primary objective is the study of the Sun’s interior through the unique tools provided by helioseismology. In recent years, the scientific scope was extended by the incorporation of instruments devoted to asteroseismology, planetary transits and earthshine monitoring. The three main small experiments are the following: • STARE (STellar Astrophysics & Research on Exoplanets) uses precise time-series photometry to search for extrasolar giant planets transiting their parent stars. An important byproduct of this search will be an unusually complete survey of variable stars within its selected fields of view. STARE is a member of TrES (Trans-Atlantic Exoplanet Survey), a network of three small-aperture telescopes searching the sky for transiting planets.s The current STARE telescope, as of July, 1999, is a field-flattened Schmidt with a nominal focal length of 286 mm and a working aperture of 99 mm, giving a focal ratio of f/2.9. The telescope is coupled to a Pixelvision 2K×2K CCD camera to obtain images with a scale of 10.8 arcseconds per pixel over a field of view 6.1 degrees square. • PASS (Permanent All Sky Survey) is an instrument for obtaining a systematic survey of the all-sky variability in stellar brightness with high temporal resolution. The project is mainly focused on the detection of planetary transits around bright stars. • EAST (Earthshine and Asteroseismology Telescope) is a stellar photometer prototype for the TEN (Taiwan Earthshine Network) project conducted by the Institute of Astronomy, National Tsing Hua University. It will be operative in the final months of 2006. Fig. 4. Three of the experiments installed at the OT. From left to right: EAST, PASS and STARE. Present and Future of Astronomy at the Observatorio del Teide 337 5 Staff The duties of the support personnel may be summed up in terms of ensuring that everything works smoothly, from the moment the astronomer first thinks of observing with any of the ITEs to obtaining the data or getting the results published. In order to improve this service, two new groups of support staff have been created in the last two years. Beginning in 2006 there is a new type of staff to support the ITEs, the Technicians in Telescopic Operations (TTOs). The TTOs are recent graduates with a training contract that to undertake work intermediate between the telescope assistants and the support astronomers. This qualified staff allows the IAC to offer a more efficient service both day and night. The other new group comprises the Professional Support Astronomers, who dedicate a minimum of a 70% of their time to support tasks. They are in charge of all the Spanish instrumentation. When possible, they are present on the first night of observation of new astronomers at the TCS, IAC80 and IACUB at OGS, as well as on the first day with TIP and TIP-II. 6 Access to the Telescopes Currently, there are two different access programmes available to support observing runs carried out at the Observatorio del Teide. Since 2004, one of these has been the OPTICON Trans-national Access Programme: Scientific, technical, logistic and financial support of observing projects submitted under the standard call launched by each facility’s TAC. This programme is available for observing time at some telescopes (THEMIS, VTT and TCS) for astronomers from EU member and associated states. It is valid for the period 2004–2008 only for a limited amount of nights per year and telescope. Another programme is the Access Programme to Relevant Scientific Facilities under the Spanish Plan for RTD and Innovation. It offers scientific, technical, logistic and financial support for observing projects submitted under the standard call launched by the Spanish TAC, responsible for the allocation of 20% of the available observing time for THEMIS and VTT and 100% for the TCS and IAC80. Through this new activities, we try to ensure that European researchers gain access to the facilities best suited to their work on the basis of scientific merit. Access is not limited by the geographic location of a particular establishment. 7 Final Remarks and Useful Links The information outlined in the previous sections aims to provide a quick and brief overview of the many things that are changing and improving at the 338 A. Oscoz Observatorio del Teide, in terms of not only new facilities and instruments but also new funding and services for astronomers. This and other information, such as the Astronomical Picture of the Month, new weather stations and a complete programme of site characterization, make the future of the Observatorio del Teide vibrant and exciting for the astronomical community. More information can be found in the following links: • • • • • • • • • • • • ITE: http://www.iac.es/telescopes/tening.html GREGOR: http://gregor.kis.uni-freiburg.de/ STELLA: http://www.aip.de/stella/ IAC80-CCD: http://www.iac.es/telescopes/iac80/CCD.htm IAC80-TCP: http://www.iac.es/telescopes/TCP/TCP.htm TCS-FIN: http://www.iac.es/telescopes/tcs/FIN.htm STARE: http://www.hao.ucar.edu/public/research/stare/stare.html OPTICON Access: http://www.otri.iac.es/opticon/ Spanish Access: http://www.otri.iac.es/na2/acceso sgic.htm Astr. Pict. of the Month: http://www.iac.es/telescopes/IAM/IndexAstrofoto ing.htm Main weather station: http://www.iac.es/telescopes/tiempo/weather.html Site Testing: http://www.otri.iac.es/sitesting/ Fig. 5. Mosaic of some of the images published in the Astronomical Picture of the Month. Prospects for the William Herschel Telescope R. Rutten Isaac Newton Group of Telescopes, Apartado de Correos 321, E-38700 Santa Cruz de la Palma, Spain, rgmr@ing.iac.es 1 abstract The development strategy for the 4.2-m William Herschel Telescope (WHT) in recent years has focused on the development of Adaptive Optics instrumentation, with the ultimate aim to provide useful high-spatial resolution imaging and spectroscopic capability for the widest possible scientific use. The scientific exploitation of Adaptive Optics (AO) with natural guide stars is severely constrained by the limited presence of bright guide stars for wavefront sensing. Use of a laser beacon as an alternative means to provide a source for wavefront sensing has the potential of drastically improving the sky coverage for AO. This paper summarized the status of the development of the GLAS laser beacon project at the WHT. Furthermore, the future exploitation of the WHT is placed in the wider context of collaboration and coordination within Europe. 2 GLAS: A Rayleigh laser guide star system for the WHT Adaptive Optics has been central to the development strategy of the 4.2-m William Herschel Telescope at the Roque de los Muchachos Observatory on the island of La Palma. The generally good seeing conditions make the site very attractive for AO exploitation. The higher spatial resolution achieved through AO provides important scientific advantages. It not only carries the advantage of distinguishing finer structure and avoiding source confusion in dense fields, but also enhances contrast of sources in background limited observations, allowing the detection of fainter objects. AO instrumentation presents an important capability for most existing large telescopes, and it is central to the development of the next generation of extremely large telescopes. But also on existing 4-m class telescopes on a good observing site AO capability 340 R. Rutten is a worth while investment and complements capability at larger telescopes by facilitating large-scale surveys at high spatial resolution. Existing AO instrumentation at the WHT is located in the Nasmyth focus of the telescope, where a purpose-built enclosure houses the AO system, NAOMI, and the various associated instruments and control systems. The NAOMI AO system is described by [2]. The AO suite of instruments include a 1024 by 1024 pixel HgCdTl infra-red camera, INGRID, an optical integral field spectrograph, OASIS, and a coronagraph unit, OSCA. In particular the OASIS spectrograph offers unique capability, and for that reason the focus of AO science exploitation lies in the wavelength range of 0.6 to 1.0 micron. Although natural guide star operation of the NAOMI AO system is now well established, the limited sky coverage has proven a serious limiting factor in its science use. For that reason in 2004 a project was embarked upon to develop a facility-class general-purpose Rayleigh laser beacon system. The project acronym, GLAS, stands for Ground-layer Laser Adaptive optics System. The overall aim of the GLAS project is to drastically improve the sky coverage for high-order AO. In order to provide the enhanced scientific capability on the shortest possible timescale, the GLAS system is relatively simple and uses commercial laser technology, without compromising the quality of engineering or safety aspects. The practical challenge of GLAS is to project a bright laser beam to an altitude of about 20 km above the observatory, bringing it to a sub-arcsecond focus, and use the Rayleigh back scattered return signal for sensing atmospheric turbulence. Fig. 1. Laser location and beam launch telescope at the top of the telescope The laser unit will be located at the top of the WHT in a temperature controlled and gravitationally stabilized unit. The laser beam, once directed Prospects for the William Herschel Telescope 341 to the beam launch telescope behind the WHT secondary mirror, will be expanded to 35cm and leave the dome, to be focused to a spot about 10 centimeters wide at a distance of 20km. Figure 1 shows a model of the laser cradle and laser beam projection telescope behind the WHT secondary mirror. The beacon’s altitude defines how well it will match the illumination of atmospheric turbulent layers of starlight coming from infinity. The mismatch between laser beacon and stars is referred to as focal anisoplanatism, or ”the cone effect” and reduces the overall performance of the AO system. Back scattered laser light will return from the whole of the atmosphere. In order to produce a tight spot for the purpose of wavefront sensing a fast shutter system is used that will admit back-scattered light to enter a very brief period after the laser pulse went out. That way only light returning from a certain distance in the atmosphere set by the light travel time will enter the wavefront sensor. The shutter will be based on Pockels cells, whose open and closed state will be set by a fast timer unit. Once having traversed the shutter light will enter a standard ShackHartmann wavefront sensor with 8x8 elements sampling the telescope pupil. The signals from this wavefront sensor will be fed back to the deformable mirror at a rate of about 300 Hz. A laser guide star AO system still requires a natural guide star for sensing the tip-tilt component of atmospheric turbulence. Since the laser light passes essentially the same atmospheric turbulence on the upward and downward pass, the return beam does not carry accurate information of the wavefront tilt induced by the atmosphere. The availability of such natural tip-tilt stars still poses a limiting factor on sky coverage, but this limitation is much reduced from the case of normal AO since the star can be much fainter. If the science object itself is bright enough to serve as tilt reference source then of course a separate star will not be required. The GLAS system will accommodate moderately extended sources such as galaxy cores for this purpose. The key aim of GLAS is to enhance sky coverage for adaptive optics observations. The resulting sky coverage will depend primarily on the chance to find a bright enough star for tip-tilt sensing. Under nominal operating conditions a star as faint as R=18 should suffice for closed loop tip-tilt sensing at 100Hz. Based on these parameters sky coverage is calculated and depicted in Figure 2 (see [3], for details). While in the case of natural guide star AO sky coverage is of order one percent or less, the GLAS system will give nearly 100% sky coverage up to a galactic latitude of 40◦ , tailing off to 50% coverage at the galactic pole. The AO correction that will be achieved at optical wavelengths will be moderate in terms of Strehl ratio, but scientifically very attractive in terms of improved FWHM. Extensive modeling has been carried out to understand the anticipated system performance. Table 1 gives a few of the key results from these model calculations, based on a laser beacon at an altitude of 20km and a natural tip-tilt star of R=17. The model includes representations of the atmosphere, telescope, laser guide star and the AO sub-systems. Performance 342 R. Rutten Fig. 2. Sky coverage diagram predictions were calculated using end-to-end Monte-Carlo simulation software developed by Richard Wilson (Durham University). The atmospheric turbulence profile was represented in 3 dimensions as a sequence of phase screens which introduce random wavefront phase aberrations with the required spatial power spectrum for atmospheric turbulence. There is good observational evidence for the localisation of atmospheric turbulence into discrete layers (eg. [4]; [1]). Here we use a representative 3 layer profile with layer strengths, heights and wind speeds chosen to match typical observed profiles for the La Palma site. The results summarized in Table 1 reflect a realistic range of seeing conditions on La Palma where median seeing is 0.6900 . Calculations were done for r◦ values of 0.11m, 0.14m and 0.19m, specified at 500nm, corresponding to natural seeing at 550nm of 0.9000 , 0.6900 and 0.5000 respectively. Image FWHM were assessed at four wavebands: R (650nm), I (850nm), J (1250nm), H (1650nm). Table 1 presents the FWHM values for the various combinations of r◦ and wavelength, for the case of a natural guide star on-axis, and 10 off-axis, covering the full range of operation. It should be noted that the results are based on statistical analysis and are therefore not exact. For reference, λ/D as a measure of the diffraction limit is shown as well. The on-axis models predict near diffraction limited performance in the J and H bands for good seeing conditions. At shorter wavelengths the diffraction limit will not be reached, and wavefront fitting errors are dominated by the performance limitations of the deformable mirror. Although Strehl ratios at short wavelengths will be low there will be a very attractive improvement in the delivered point spread function. It should be noted that any source as faint as R=17, and possibly even fainter, may serve as a self-referencing tip-tilt source. Hence for any such point source good AO correction will be obtained. Prospects for the William Herschel Telescope 343 Table 1. GLAS performance expectations; FWHM in arcseconds for different values of r◦ and wavebands Natural tip-tilt star on-axis Natural tip-tilt star 10 off-axis λ/D r◦ 0.19 0.14 0.11 0.19 0.14 0.11 R 0.12 0.27 0.41 0.18 0.31 0.45 0.03 I 0.09 0.17 0.29 0.14 0.24 0.35 0.04 J 0.09 0.13 0.18 0.13 0.19 0.24 0.06 H 0.10 0.12 0.16 0.12 0.16 0.20 0.08 For fainter science targets or diffuse sources an off-axis natural guide star is required, causing some degradation of AO performance. From the actual performance of the NAOMI AO system we expect these model calculations to be realistic. Based on the La Palma seeing statistics we anticipate that suitable seeing conditions for general AO and laser beacon operation will exist about 75% of the available observing time ([5]). Periods for which ground layer turbulence dominates that hence will be especially suitable for GLAS occur about 25% of the time. On occasions scattering by dust or thin cirrus in the atmosphere may hamper efficient laser operation, but this is expected to interfere with observations only a small fraction of the time. The operating mode of the WHT will shift towards queue scheduled observing in order to exploit the best possible atmospheric conditions for those top science proposals that require the best seeing conditions, bolstering the science return of the telescope. 3 The WHT in the European context The home of the William Herschel Telescope and of several other world-class astronomical facilities is the Roque de los Muchachos Observatory on the island of La Palma, Europe’s prime observing site in the Northern hemisphere. The developement of the 10-m GTC underlines this fact. The combination of very good observing conditions and a well developed infrastructure makes it an attractive site for many types of ground-based astronomical observations. The advent of the latest generation of very large telescopes on the one hand, and the development of dedicated medium-size telescopes on the other hand, is rapidly changing the role of the existing 2-m to 4-m class telescopes. The scientific function of these medium-size telescopes is becoming ever more focused on specific functionality and scientific projects, driven by scientific needs as well as the need to reduce operational costs. This changing role of existing medium-size telescopes, together with mounting financial pressures suggests that coordination at a wider European level would be advantageous 344 R. Rutten to astronomy in Europe. The observatory on La Palma, and the Spanish astronomical community, could play a leading role in such an initiative. References 1. J.J. Fuensalida, B. Garcı́a-Lorenzo, J. Castro, S. Chueca, J.M. Delgado, J.M. González-Rodrı́guez, C. Hoegemann, M. Reyes, M. Verde, J. Vernin: 2004, Proc. Remote Sensing (in press) 2. R.M. Myers, R.M., Longmore, A.J., Benn, C.R., et al., 2003, Proc. SPIE, Vol 4839, 647 3. Stuik, R., LeLouarn, M., Quirrenbach, A., 2004, Proc. SPIE, Vol 5490, 331 4. Vernin, J., Muñoz-Tuñón, C., 1994, A&A 284, 311 5. Wilson, R.W., O’Mahony, N., Packham, C., Azzaro, M., 1999, MNRAS 309, 379 VO Science. The Spanish Virtual Observatory E. Solano INTA-LAEFF. Apartado de Correos 50727. 28080 Madrid, Spain, esm@laeff.inta.es Summary. After a brief description of the reasons that drove the creation of the Virtual Observatory I will focus on the role that science has in the VO projects and the impact that using a VO methodology will have on it. Finally, I will describe the Spanish Virtual Observatory and some of the initiatives that are being developed in its framework. 1 The Virtual Observatory The advances in technology (telescope design and fabrication, large-scale detector arrays, highly automated reduction pipelines,...) are now permitting to collect multi-parameter information of large regions of the sky. In addition to this, advances in computational capabilities are providing the means to make, for the first time, direct comparisons between complex theoretical calculations and large, statistically significant observational databases. All these data are easily accessible from user-friendly, on-line information systems that are intensively used by the astronomical community. Although this scenario of multi-Terabyte datasets interlinked through high-speed networks should lead to a more complete and less biased understanding of complex astrophysical phenomena, the reality is that the progress in the scientific exploitation is not keeping pace with the exponential growth of data. The reason for this is mostly related to the methodology normally used in astronomical research: once the data of interest have been identified, they are transferred to the astronomer’s desktop for their reduction and analysis using standard astronomical software. Even if this methodology has proven to be valid in many cases, it has demonstrated to be quite inefficient when dealing with problems that require either interoperability among data services (a basic issue for multiwavelength astronomy) and/or management of large volumes of data. The Virtual Observatory (VO) is a complementary approach to this classical methodology. In the same manner as the Web constitutes a distributed, 346 E. Solano interlinked system providing easy access to information all over the world, the final objective of the Virtual Observatory is that all the world’s astronomical data should feel like it sits on the astronomer’s desk top. Phase-A of the main VO projects (NVO, AVO, Astrogrid) began in 2001. In June 2002, IVOA1 was created as the framework to coordinate these and other similar initiatives. At present, IVOA is formed by 16 projects. In Europe the Virtual Observatory initiatives are being built around EURO-VO. Based on the experience gained within the AVO project, EURO-VO is conceived as a Phase-B project with the objective of deploying a fully operational VO in Europe. Further information about EURO-VO can be found in [6]. Similar transition activities in other VO projects can be found in [2]. 1.1 Interoperability among astronomical services Interoperability is a key requirement to provide the scientific community with an integrated access to the wide variety of information of very diverse nature (astronomical data in image or tabular form, catalogues, bibliographic information, ...) that is available from astronomical services. Although most of these services have implemented user-friendly interfaces allowing easy retrieval of their contents, the different formats, units, policies, retrieval protocols,..., makes it very inefficient the data compilation from more than one archive. In practice, the simultaneous query to multiple databases is done, most of the times, slowly and painfully by hand, a situation that represents a major drawback for multiwavelength analysis. The Virtual Observatory proposes a three-step solution for this problem: agreement and development of standards, uptake of these standards by the data services and development of the federation of astronomical data centres (“data grid”). Standardization is, therefore, the key concept to solve the interoperability problems: Semantic, access protocols, data formats, data models and registries are some of the fields where the Virtual Observatory Alliance is working in defining standards (more information on these topics can be found at the IVOA portal). 1.2 Management of Large Volumes of Data Vast amount of information is being continuously incorporated to astronomical archives. Large-area sky surveys like SDSS, 2MASS and DENIS have associated millions of observations. Planned survey-dedicated telescopes like VISTA, VST or LSST, with a data rate a thousand times larger, will even make things worse ending up in a situation in which the classical approach of transferring data from the archive to the local desktop for their further analysis will be, by far, not applicable. 1 http://www.ivoa.net VO Science. The Spanish Virtual Observatory 347 The solution provided by the Virtual Observatory to solve this problem is based on the service paradigm. According to this approach, data will be left as it resides and only results will be delivered to users. This scenario puts strong requirements on the server side as it must be able to manage intensive processing and analysis tasks using either local resources or distributed computing like the Grid. 1.3 Analysis tools. Data mining The potential for scientific discovery afforded by the mass of data available from the Virtual Observatory services is enormous: New and unexpected scientific results of major significance will undoubtedly emerge from the combined use of datasets and services, a type of science that would not be possible from such sets used singly. However, transforming vast masses of bits into a refined knowledge and understanding of the universe is a highly complex task that requires very specific analysis tools, running remotely but with a performance comparable to that obtained if they were local (see, for instance, IVOA2 for a detailed list of VO applications). Data mining, defined as the non-trivial extraction of previously unknown implicit information, represents the best way to perform an efficient and systematic study of the vast amount of VO data. Automated classification based on observations, the discovery of the unknown using unsupervised methods, the search of rules and properties, ...., are some of the fields that clearly benefit from the use of data mining techniques. 2 Science in the Virtual Observatory projects Although technologically enabled, the Virtual Observatory must not be seen as a mere technological project since it is driven by science as its final goal is to produce better, new and more efficient science. This concept was clearly understood by the major VO projects which, since very early, set up Science Working Groups to provide scientific advice to the projects. One of the major tasks of these groups was the identification, with a clear emphasis on the definition of the science requirements, of cases that could benefit from the use of the VO technology. The ScienceProblem3 or the TopTen4 lists compiled by AVO or AstroGrid, consisting of a number of key science cases that should be achieved once the EURO-VO is fully operational, are examples of this. Nowadays, the Virtual Observatory has already reached a level of maturity enough to produce relevant science results. The first major discovery made with the Virtual Observatory is described in [7]: the AVO science 2 3 4 http://www.ivoa.net/twiki/bin/view/IVOA/IvoaApplications http://wiki.astrogrid.org/bin/view/VO/ScienceProblemList http://wiki.astrogrid.org/bin/view/Astrogrid/ScienceProblems 348 E. Solano team involved in this project discovered 31 previously undetected supermassive black holes in the GOODS fields following a VO methodology. Since then, several new scientific findings through VO-based research have been already published. The classification of ROSAT sources using multiwavelength information and data mining techniques ([4]) or the discovery of massive, dust-enshrouded, carbon stars ([10]) are good examples of this. A review of the new (present and coming) VO Science took place during the meeting5 held in August 2006 in the framework of the General Assembly of the International Astronomical Union. Eighty contributions (53 oral) covering from the Sun-Earth connection to Cosmology demonstrated the growing interest of the international astronomical community about the Virtual Observatory. 3 The Spanish Virtual Observatory (SVO) The Spanish Virtual Observatory is a project funded by the Ministry of Education and Research that officially joined IVOA in June 2004. Soon, it was clear that the project was not enough to cover the different initiatives of the Spanish community. This led us to apply for a Thematic Network associated to the project whose primary goal is to enhance the collaboration among the Spanish groups with interest in VO through the organization of meetings, tutorials and workshops taking advantage of the potential synergies that may arise. One of these activities is the SVO school that will be held in November 2006: Open to anyone interested in learning how to use the VO for astronomical research, to become familiar with the high performance capabilities of the Virtual Observatory, to develop VO-aware tools, or to make astronomical data collections available to VO users, the school will have a twofold objective: Identify the potential benefits that the application of a VO methodology may produce and develop VO-enabled research project. At present, the network is formed by 84 members from more than 20 institutes and departments with very diverse interests. Among them, I would stress the following: • • • • • 5 6 Take-up of VO standards (archives and catalogues) Development of VO services VO compliance of theoretical models Data mining tools VO Science More information about these topics can be found in the SVO portal6 . http://www.ivoa.net/pub/VOScienceIAUPrague/programme/ http://svo.laeff.inta.es VO Science. The Spanish Virtual Observatory 349 3.1 SVO Services: VOSED, a tool for the characterisation of Circumstellar Disks Knowledge of the properties of protoplanetary disks and how they evolve from the dense and turbulent disks around PMS stars to the tenuous debris disks around MS objects is a crucial step to understand the formation of extrasolar planetary systems. This analysis requires studying a statistically significant number of stars covering a wide range of masses and evolutionary stages (HAeBes, ETTs, CTTs, PTTs, A-shell and Vega-type stars) to track and characterize from an observational point of view the evolution of protoplanetary disks. Theoretical modelling of SEDs has proven to constitute an invaluable tool for understanding the structure and properties of protoplanetary disks. However, SEDs building requires accessing to a variety of astronomical services providing, in most of the cases, heterogeneous information. Moreover, model fitting and the associated detailed analysis demands a tremendous amount of work and time which makes it very inefficient even for a modest dataset. In the framework of the Spanish Virtual Observatory we are developing an application to build observed spectral energy distributions taking advantage of the already existing VO standards and tools. The application, called VOSED7 , queries the VO registry to gather all the spectroscopic information available for the object of interest, information that is later complemented with queries to other VO services like, for instance, VizieR. Model fitting constitutes a fundamental step in the analysis process. VOSED includes a tool to estimate the model parameters (both stellar and disk) following a Bayesian method. The main aim of the tool is to quantitatively analyse the data in terms of the evidence of models of different complexity, evaluate what other alternative models can compete with the most a posteriori probable one and what are the most discriminant observations to discard alternatives. Preliminary analysis have proven the ability of the system to reproduce the results obtained by [5] in a much more efficient and homogeneous way (Figure 1). 3.2 SVO Data Mining: Automated classification of light curves Both ground-based projects (ASAS, OGLE, MACHO,...) and space missions (INTEGRAL, COROT, KEPLER in the near future and GAIA in a medium term) are producing photometric time series for a huge number of objects, many of them observed for the first time, and for which human-based classification systems are, by far, not applicable. In this context, SVO has developed a tool for the detection and classification of light curves of both eclipsing binary systems ([9]) and periodic variable stars ([3]) using different methodological strategies like neural networks, support vector machines and bayesian 7 http://sdc.laeff.esa.es/vosed 350 E. Solano Fig. 1. HD 34282 Spectral Energy Distribution fitting using VOSED. network classifiers. A preliminary version of the classification system is already implemented in the OMC archive8 (Figure 2). A new version will be released in February 2007 and will constitute the basis of the classification system adopted for the COROT mission. 3.3 SVO Science. Searching for Rare Objects: Brown Dwarfs A fundamental question in the area of star formation is the form of the stellar mass function at the lower end, in particular, the contribution of brown dwarfs to the mass budget. Brown dwarfs are objects occupying the gap between the 8 http://sdc.laeff.inta.es/omc/ VO Science. The Spanish Virtual Observatory 351 Fig. 2. The automated classification system implemented in the OMC archive. least massive stars and the most massive planets. They are intrinsically faint objects so their detection is not straightforward. The advent of global surveys at deep optical and near-infrared wavebands like 2MASS, SDSS or DENIS allowed mining the sky through an appropriate combination of colours and/or proper motion information (since any detectable brown dwarf must be nearby and, so, it will have relatively high proper motion) alleviating, at least partly, this situation. Nevertheless, building a complete census of substellar objects implies the discovery of a statistically significant number of them through queries that combine attributes available from different archives, an approach out of the scope of the “classical” research methodology but that perfectly fits in the framework of the Virtual Observatory as demonstrated by the discovery of a new L-type brown dwarf during the 2003 NVO Science Demonstration. Using 2MASS, DENIS and SDSS and Aladin as VO-tool, we are focusing on the nearby (d ≤ 10 pc) population of brown dwarfs with T spectral type. A similar methodology will be apply for the exploitation of UKIDSS, a near-IR survey three magnitudes deeper than 2MASS and whose first Data Release took place in July 2006. With UKIDSS it will be possible to discover faint T dwarfs beyond the 2MASS limiting magnitude and to proof the existence of the theoretically predicted Y-type brown dwarfs, a class of objects cooler than the T dwarfs and whose discovery constitutes one of the key science drivers for the UKIDSS project. We are also use VO tools to try to shed light on the problem of the brown dwarf formation, an issue still open. Among the competing theories, [8] proposed that brown dwarfs are stellar embryos ejected from the formation region 352 E. Solano due to gravitational interactions before their hydrostatic cores could build up enough mass to eventually start hydrogen burning. Our goal is to check this model by cross-correlating IPHAS ([1]) and 2MASS to search, for the first time, young brown dwarfs over large areas of the sky on the basis of their Hα emission and IR colors. Preliminary analysis of the follow-up observations of some of the candidates seems to confirm the existence of several young brown dwarf among them. References 1. J. E. Drew, R. Greimel, M. J. Irwin et al: MNRAS 362, 753 (2005) 2. R. Hanisch: The Virtual Observatory in Transition. In: ASP Conf. Ser., vol 351, ed by C. Gabriel, C. Arviset, D. Ponz, E. Solano (ASP, San Francisco 2006) pp 765–770 3. M. López, C. Bielza, L. M. Sarro: Bayesian classifiers for variable stars. In: ASP Conf. Ser., vol 351, ed by C. Gabriel, C. Arviset, D. Ponz, E. Solano (ASP, San Francisco 2006) pp 161–164 4. T. A. McGlynn, A. A. Suchkov, E. L. Winter, E. L. et al: ApJ 616, 1284 (2004) 5. B. Merı́n, B. Montesinos, C. Eiroa, et al: A&A 419, 301 (2004) 6. P. Padovani: The AVO to EURO-VO Transition. In: ASP Conf. Ser., vol 351, ed by C. Gabriel, C. Arviset, D. Ponz, E. Solano (ASP, San Francisco 2006) pp 771–774 7. P. Padovani, M. G. Allen, P. Rosati et al: A&A 424, 545 (2004) 8. B. Reipurth, C. Clarke: AJ 122, 432 (1995) 9. L. M. Sarro, C. Sánchez-Fernández, A. Giménez: A&A 446, 295 (2005) 10. P. Tsalmantza, E. Kontizas, L. Cambrsy et al. A&A 447, 89 (2006) A Appendix: Table of Contents of the CD-Rom Session I. Galaxies and cosmology Contributed talks A numerical 3+1 code for GRMHD L. Antón, J.A. Miralles, J.M. Martı́, J.M. Ibáñez Astrophysical Effects of (Light) Dark Matter Y. Ascasibar The formation epoch of the galaxy red sequence M. Balcells A FIRST-APM-SDSS Survey for High-Redshift Radio QSOs R. Carballo, J.I. González-Serrano, F.M. Montenegro-Montes, C.R. Benn, K.-H. Mack, M. Pedani, M. Vigotti The XMM-Newton and Spitzer view of galaxy/AGN formation at z=2-3 F.J. Carrera, J. Ebrero, M. Page, J. Stevens Study of local Luminous Compact Blue Galaxies based on Integral Field Spectroscopy A. Castillo-Morales, J. Pérez-Gallego, J. Gallego, R. Guzmán Kinematics of globular clusters in NGC 1407: constraining the galaxy mass and its dark matter halo A.J. Cenarro, M.A. Beasley, J. Strader, J.P. Brodie, D.A. Forbes Properties in the u band of local star-forming galaxies C. Dı́az, J. Zamorano, J. Gallego, P.G. Pérez-González, A. Gil de Paz Reconstructing Weak and Strong Lensing in Galaxy Clusters J.M. Diego Flux distribution and angular clustering of sources in the X-ray Universe J. Ebrero, F.J. Carrera Principal Component Analysis as a tool to explore star formation histories I. Ferreras, B. Rogers, O. Lahav Extended UV Emission in Nearby Spiral Galaxies A. Gil de Paz, B.F. Madore, D. Thilker, S. Boissier, L. Bianchi, the GALEX Team Gravitationally Lensed QSOs: Optical Monitoring with the EOCA and the Liverpool Telescope (LT) L.J. Goicoechea, A. Ullán, J.E. Ovaldsen, E. Koptelova, V.N. Shalyapin, R. GilMerino The Growth of Dark-Matter Halos and their Typical Structural and Kine- 354 A Appendix: Table of Contents of the CD-Rom matic Properties G. González-Casado, E. Salvador-Solé, A. Manrique Early-type galaxies colour gradient and their relation with global galactic properties V. González-Pérez, F.J. Castander, G. Kauffmann On the accuracy in the derivation of elemental abundances of the emitting gas in HII galaxies G.F. Hägele, A.I. Dı́az, E. Pérez-Montero, E. Terlevich, R. Terlevich Environment of compact extragalactic radio sources A. Labiano, C.P. O’Dea, P.D. Barthel, R.C. Vermeulen The FIR properties of the most isolated galaxies U. Lisenfeld, L. Verdes-Montenegro, J. Sulentic, S. Leon, D. Espada, G. Bergond, E. Garcı́a, J. Sabater, J.D. Santander-Vela, S. Verley Massive star formation in Wolf-Rayet galaxies A.R. López-Sánchez, C. Esteban Study of the CO index at 2.3 µm: a new stellar library in the K band E. Mármol-Queraltó, N. Cardiel, A.J. Cenarro, A. Vazdekis, J. Gorgas, R.F. Peletier Evolutionary self-consistent models of HII galaxies M. Martı́n Manjón, M. Mollá, A.I. Dı́az Nuclear Activity in UZC Compact Groups of Galaxies M.A. Martı́nez, A. del Olmo, P. Focardi, J. Perea Abundance-dependent cooling functions for low-density astrophysical plasmas in equilibrium F.J. Martı́nez-Serrano, A. Serna, R. Domı́nguez-Tenreiro Local Group candidates in cosmological simulations L.A. Martı́nez, G. Yepes, M. Sivan, Y. Dover, Y. Hoffman AGN population studies with the 2XMM X-ray source catalogue S. Mateos, M.G. Watson, J.A. Tedds, the XMM-Newton Survey Science Centre Evolutionary synthesis models for spirals and irregular galaxies M. Mollá Specific SFR profiles in nearby spiral galaxies: quantifying the inside-out formation of disks J.C. Múñoz-Mateos, A. Gil de Paz, S. Boissier, J. Zamorano, T. Jarrett, J. Gallego, B.F. Madore Bright and Dark Matter in Elliptical Galaxies J. Oñorbe, R. Domı́nguez-Tenreiro, A. Sáiz, H. Artal, A. Serna Baryonic Matter at Supercluster Scales: the case of Corona Borealis Supercluster C.P. Padilla-Torres, R. Rebolo, C.M. Gutiérrez, R. Génova-Santos, J.A. Rubiño, R. Watson The radio and X-ray properties of low-luminosity active galactic nuclei F. Panessa, X. Barcons, L. Bassani, M. Cappi, F.J. Carrera, M. Dadina, L.C. Ho, S. Pellegrini Spitzer insight into galaxy evolution in the last 12 Gyr (from z˜4) P.G. Pérez-González, the MIPS Instrument Team Spectrophotometry of the polar ring galaxy IIZw71 E. Pérez-Montero, R. Garcı́a-Benito, A.I. Dı́az The Narrow-Line Region of the Seyfert 2 galaxy Mrk 78: An Infrared A Appendix: Table of Contents of the CD-Rom 355 View C. Ramos Almeida, A.M. Pérez Garcı́a, J.A. Acosta Pulido, J.M. Rodrı́guez Espinosa, R. Barrena, A. Manchado On the distribution of polarization directions in quasar samples D. Sáez, J.A. Morales Morphologies and stellar populations of galaxies in the core of Abell 2218 S.F. Sánchez, N. Cardiel, M.A.W. Verheijen, S. Pedraz, G. Covone Deep hi-Observations and Tully-Fisher Distances of Six Spirals in the Virgo Cluster Region M.C. Toribio J.M. Solanes Extragalactic star clusters in Merging/ Interactive Galaxies G. Trancho, N. Bastian, F. Schweizer, B.W. Miller Dust and Dark Matter in a Giant Elliptical Galaxy at an Intermediate Redshift A. Ullán, L.J. Goicoechea, R. Gil-Merino Completion of the GOYA Photometric Survey M. Vallbé i Mumbrú, M. Balcells, D. Abreu, G. Barro, D. Cristóbal, L. Domı́nguez, C. Eliche, J. Gallego, C. López, M. Prieto, V. Villar Maximum Oxygen Abundance in Spiral Galaxies and the MetallicityLuminosity Relation J.M. Vı́lchez, L.S. Pilyugin, T.X. Thuan Star formation on an Hα selected sample at z=0.84 V. Villar, J. Gallego, S. Pascual, J. Zamorano, K. Noeske, D.C. Koo Posters Applications of Volumetric Visualization Methods in Hydrodynamical Cosmological Simulations H. Artal, J. Oñorbe, R. Domı́nguez-Tenreiro, F.J. Martı́nez-Serrano, L. Benjouali Analysis of the brightest galaxies in the central part of clusters of galaxies B. Ascaso Anglés, M. Moles Villamate, J.A.L. Aguerri, R. Sánchez-Janssen, J. Varela The Orbits of Disk Galaxy Satellites from Cosmological Hydrodynamical Simulations: How Populated Are Polar Trajectories? L. Benjouali, M.A. Gómez-Flechoso, H. Artal, R. Domı́nguez-Tenreiro Short-Period Variables in the Local Group Dwarf Galaxies Tucana and LGS3 E.J. Bernard, M. Monelli, C. Gallart, the LCID Team: A. Aparicio, G. Bertelli, S. Cassisi, A.A. Cole, P. Demarque, A.E. Dolphin, I. Drozdovsky, H.C. Ferguson, S. Hidalgo, M. Mateo, L. Mayer, J. Navarro,1, F. Pont, E.D. Skillman, P.B. Stetson, E. Tolstoy Relationship between the optical and X-ray characteristics of extragalactic sources J. Bussons Gordo, X. Barcons, F.J. Carrera, M.T. Ceballos, A. Corral, J. Ebrero, S. Mateos, M.J. Page, M.G. Watson Soft X-ray spectra of Seyfert 1 galaxies observed with XMM-Newton M.V. Cardaci, M. Santos-Lleó, A.I. Dı́az Diffuse light in the Virgo cluster 356 A Appendix: Table of Contents of the CD-Rom N. Castro-Rodrı́guez, J.A.L. Aguerri The origin of the brightest half of the X-ray power generated in the Universe: XMS sample presentation M.T. Ceballos, X. Barcons, F.J. Carrera, S. Mateos, A. Corral, J. Bussons, M.J. Page, M.G. Watson The OTELO Project J. Cepa, E.J. Alfaro, H.O. Castañeda, J. Gallego, J.I. González-Serrano, J.J. González, D. Heath Jones, A.M. Pérez-Garcı́a, M. Sánchez-Portal The 2dF/ XMM-Newton Survey A. Corral, S. Mateos, F.J. Carrera, M.J. Page, X. Barcons, M.T. Ceballos, M. Watson, J. Tedds Astroparticle Physics at the IAC M.T. Costado on behalf of the Astroparticle Physics Group ofthe IAC The jet-cloud interaction in 3CR galaxies: The case of 3C 135, 3C 80, 3C 234, 3C 284 C. Feinstein, F.D. Macchetto, M.F. Montero, G.F. Hägele Results of the first year of extragalactic observations with MAGIC R. Firpo, for the MAGIC collaboration Radiosources in the OTELO survey: Groth field J.I. González-Serrano, A.M. Pérez-Garcı́a, M. Sánchez-Portal, J. Cepa, J. Gallego, H. Castañeda, E.J. Alfaro, J.J. González A spectrophotometric study of the physical parameters of circumnuclear star forming regions G.F. Hägele, A.I. Dı́az, M.V. Cardaci, E. Terlevich, R. Terlevich, M. Castellanos HII Galaxies with 3D Spectroscopy C. Kehrig, J.M. Vı́lchez, S.F. Sánchez, E. Telles, D. Martin-Gordón, L. LópezMartı́n Metallicity-dependent chemical feedback in cosmological SPH simulations F. J. Martı́nez-Serrano, A. Serna, M. Mollá, R. Domı́nguez-Tenreiro Dynamical and rotational masses and spectroscopic properties of a sample of Hα-selected galaxies at z=0.24 S. Pascual, J. Gallego, J. Zamorano An XMM-Newton study of Hyper-Luminous Infrared Galaxies A. Ruiz, F.J. Carrera, F. Panessa X-ray emitters in OTELO survey: the Groth Strip M. Sánchez-Portal, A.M. Pérez-Garcı́a, J. Cepa, J.I. González-Serrano, E.J. Alfaro, H. Castañeda, J. Gallego, J.J. González, the OTELO collaboration Supercomputer Simulations of Compact Groups of Galaxies: the IDILICO project J.M. Solanes, E. Athanassoula, A. Bosma, C. Garcı́a-Gómez, J.C. Lambert, A. del Olmo, J. Perea, M.C. Toribio High Resolution Spectroscopy Of Dwarf Elliptical Galaxies Within The MAGPOP-ITP collaboration E. Toloba, J. Gorgas, A.J. Cenarro, A. Bosseli, N. Cardiel, C. Carretero, A. Gil de Paz, E. Mármol-Queraltó, D. Michielsen, R.F. Peletier, I. Pérez, A. Vazdekis Light curve analysis of eclipsing binaries and Cepheids in the Andromeda galaxy F. Vilardell, C. Jordi, I. Ribas A Appendix: Table of Contents of the CD-Rom 357 Session II. The Galaxy and its components Contributed talks Fundamental astrometry and experiments in gravitation with Gaia using the planets of the Solar System G. Anglada-Escudé, S. Klioner, J. Torra Neutrino-driven wind in supernova cores A. Arcones, L. Scheck, H.T. Janka Broad-band emission from high-energy processes in microquasars V. Bosch-Ramon, J.M. Paredes The Magellanic Clouds Chemical Enrichment History and Its Gradient R. Carrera, C. Gallart, A. Aparicio, E. Costa, E. Hardy, R. Méndez, E. Pancino The initial-final mass relationship of white dwarfs in common proper motion pairs S. Catalán, I. Ribas, J. Isern, E. Garcı́a-Berro, C. Allende Prieto Studies of SiO circumstellar maser emission with very high spatial resolution F. Colomer, R. Soria-Ruiz, J. Alcolea, V. Bujarrabal, J.F. Desmurs Kinematic evolution of the young local associations and the Sco-Cen complex D. Fernández, F. Figueras, J. Torra Hot Massive Stars in the Far-UV and UV ranges M. Garcı́a The star formation history of NGC 5471 R. Garcı́a-Benito, E. Pérez, A.I. Dı́az The abundance discrepancy problem in H II regions J. Garcı́a-Rojas, C. Esteban Metallicity distribution of the inner parts of the Milky Way: a spectroscopic follow-up to TCS-CAIN C. González-Fernández, A. Cabrera-Lavers, P.L. Hammersley, F. Garzón Positrons in Type Ia Supernovae A. Hirschmann, E. Bravo, J. Isern Diffuse bands, extinction and anomalous microwave emission in the Interstellar medium:the role of fullerenes S. Iglesias-Groth The First Measurements of the Electron Density Enhancements Expected in C-shocks I. Jiménez-Serra, J. Martı́n-Pintado, S. Viti, S. Martı́n, A. Rodrı́guez-Franco, A. Faure, J. Tennyson Galactic Warp explains the overdensity of the Canis Major Region M. López-Corredoira Stellar spectra parametrization by neural networks: extraction of Teff, logg and [F e/H] in the Gaia RVS spectral region M. Manteiga Outeiro, D. Ordóñez Blanco, C. Dafonte Vázquez, B. Arcay Varela, I. Carricajo Marı́n Small young open clusters with unusual upper IMF 358 A Appendix: Table of Contents of the CD-Rom A. Marco, I. Negueruela Wavelength-dependence of the SN 1993J expansion rate confirmed I. Martı́-Vidal, J.M. Marcaide, A. Alberdi High Mass X-ray Binaries seen by INTEGRAL S. Martı́nez-Núñez, P. Blay, A. Camero Arranz, P. Reig, P.H. Connell, D. Miralles, V. Martı́nez Badenes IRS 8: An Outsider in the Galactic Center F. Najarro, T.R. Geballe Interaction and on-going star formation in a closely-packed environment A. Palau, R. Estalella, J.M. Girart, P.T.P. Ho, Q. Zhang, H. Beuther OTELO: The stellar component of the Groth field A.M. Pérez Garcı́a, E.J. Alfaro, J. Cepa, A. Bongiovanni, H. Castañeda, J. Gallego, J.J. González, J.I. González Serrano, M. Sánchez-Portal, the OTELO collaboration EVN+MERLIN wide-field imaging of the high-redshift radio galaxy 4C 41.17 M.A. Pérez Torres Circumstellar Reddening in Be/X-ray Binaries of the Magellanic Clouds and The Galaxy M. S. Riquelme, J.M. Torrejón Evidence of a possible second cyclotron line in the X-ray spectrum of 4U 1538-52 J.J. Rodes, J.M. Torrejón G. Bernabéu The formation of rings and spirals in barred galaxies M. Romero-Gómez, E. Athanassoula, J. Masdemont, C. Garcı́a-Gómez XMM-Newton shows the highly ionized disk wind of the microquasar GRO J1655-40 G. Sala, J. Greiner, J. Vink, F. Haberl Interaction between the ejected mass in supernovae explosions with different types of secondary stars N. Serichol, D. Garcı́a-Senz Discovery of microquasar LS I +61 303 at VHE gamma-rays with MAGIC N. Sidro, for the MAGIC collaboration Regions of star formation as γ-ray sources D.F. Torres CdC-SF Catalogue: An Astrometric Catalogue using emphCarte du Ciel plates, San Fernando zone B. Vicente, C. Abad, F. Garzón Posters Use of the Besanccon Galaxy Model for the Gaia Mission Simulator E. Amôres, A. Robin, X. Luri, E. Masana, C. Reylé, C. Babusiaux Dominant mechanisms in the evolution of kinematic structures T. Antoja, F. Figueras D. Fernández Evolution of stable and unstable Galactic disks A. Artigas, A. Burkert, J. Isern, E. Bravo uvby-Hβ CCD photometry and membership segregation of the open cluster NGC 2682 (M 67) L. Balaguer-Núñez, C. Jordi, D. Galadı́-Enrı́quez A Appendix: Table of Contents of the CD-Rom 359 Microquasar jets interacting with their environments P. Bordas, J.M. Paredes, V. Bosch-Ramon Multiple YSOs in the Low-Mass Star-Forming Region IRAS 00213+6530 G. Busquet, R. Estalella, A. Palau, J. M. Girart, G. Anglada, I. Sepúlveda ToO observations of GRO J1655-40 in outburst with INTEGRAL M.D. Caballero-Garcı́a, E. Kuulkers, P. Kretschmar, A. Domingo, J.M. Miller, J.M. Mas-Hesse Refractive Index of Ices in the UV-Vis Range J. Cantó, C. Millán, M. Domingo, M.A. Satorre Measuring line-strength indices in a systematic way N. Cardiel Blue massive stars in NGC55 and IC1613 N. Castro, M. Garcia, A. Herrero, C. Trundle, F. Bresolin, A. Rosenberg, L. Corral, R. Demarco, I. Flores Nova Oph 1998 observed with XMM-Newton C. Ferri, M. Hernanz, G. Sala Lunar Occultations with VLT-ISAAC set a new standard for high angular resolution, high-sensitivity measurements O. Fors, A. Richichi, E. Mason, J. Stegmeier, T. Chandrasekhar NIR Observations with LIRIS of Core-Collapse Supernovae G. Gómez, R. López, J.A. Acosta-Pulido XMM-Newton RGS Detection of Emission Lines of O VII and N VI in the X-ray Spectrum of the Cat’s Eye Nebula M.A. Guerrero, Y. Chu, R.A. Gruendl Active regions in a sample of late-type rotator stars: spectroscopic analysis M. Hernán-Obispo, E. de Castro, M.C. Gálvez, M. Cornide, D. Montes Observations of CMB and Galactic Foregrounds with the COSMOSOMAS experiment. Evidence for spinning dust S.R. Hildebrandt Gaia: photometric data treatment C. Jordi, J.M. Carrasco, E.B. Amôres, C. Fabricius, F. Figueras A Survey for Brown Dwarfs in Corona Australis and Chamaeleon II B. López Martı́, J. Eislöffel, R. Mundt Integral Field Spectroscopy of Herbig Haro Objects: HH 110 R. López, S.F. Sánchez, G. Gómez, B. Garcı́a-Lorenzo, R. Estalella, A. Riera Spectral atlas of HH 262 R. López, S.F. Sánchez, B. Garcı́a-Lorenzo, G. Gómez, A. Riera, R. Estalella The Long Bar in the Milky Way M. López-Corredoira, A. Cabrera-Lavers, F. Garzón, T.J. Mahoney, P.L. Hammersley, C. González-Fernández The Young Stellar Populations in the Solar Neighbourhood J. López-Santiago, G. Micela, S. Sciortino, F. Favata A kinematical study of the irradiated dense core ahead of HH 80N J.M. Masqué, J.M. Girart, R. Estalella Temperature scale for FGK giant stars E. Masana, C. Jordi SMC: Clues about the star formation history N.E.D. Nöel, C. Gallart, E. Costa, R.A. Méndez 360 A Appendix: Table of Contents of the CD-Rom High Energy Sources Observed with OMC D. Rı́squez, A. Domingo, J.M. Mas-Hesse, E. Kuulkers Spectroscopy of a giant star formation region in NGC 6946 M. C. Sánchez-Gil, E.J. Alfaro, E. Pérez Centimeter Emission from Selected High-Mass Star-Forming Regions A. Sánchez-Monge, A. Palau, R. Estalella, M.T. Beltrán, J.M. Girart Eclipses of X-rays Flares in Binary Stars J. Sanz-Forcada, F. Favata, G. Micela First results of galactic observations with MAGIC N. Sidro, for the MAGIC collaboration Contribution of different stellar populations to the galactic content of 26 Al and 60 Fe M. Suades, M. Hernanz, N. de Séréville, G. Martı́nez-Pinedo, J. José Session III. The Sun and planetary systems Contributed talks Pleiades low-mass brown dwarfs G. Bihain The effect of ultraviolet irradiation of mixture ice rich in CO2 J. Cantó, M.A. Satorre, R. Luna, C. Santonja A high resolution spectrograph for the 1.5m solar telescope GREGOR M. Collados, J.J. Dı́az, E. Hernández, R. López, E. Páez Relationships between physical and observational parameters during flares on M-dwarfs I. Crespo-Chacón, D. Garcı́a-Alvarez, D. Montes, M.J. Fernández-Figueroa, J. López-Santiago, D. Jevremovic, B.H. Foing Microphysics of Titan’s methane storms R. Hueso, A. Sánchez-Lavega Numerical simulations of magneto-acoustic waves in a sunspot-like magnetic structure E. Khomenko, M. Collados Atmospheric Turbulence in the Clouds of Venus J. Peralta, R. Hueso, A. Sánchez-Lavega Possible Signatures of Different Formation Mechanisms in the EccentricityMass Distribution of Exoplanets I. Ribas, J. Miralda-Escudé Effect of the presence of carbon dioxide on the Triton’s icy surface M.A. Satorre, J. Cantó, M. Domingo, R. Luna, C. Millán C. Santonja Several bolides studied by the Spanish Fireball Network: Villalbeto de la Peña and other 2003-2006 events J.M. Trigo-Rodrı́guez, J. Llorca, J.L. Ortiz, J.A. Docobo, A.J. Castro-Tirado Posters A Appendix: Table of Contents of the CD-Rom 361 Monte Carlo simulations applied to solar electron events. A case study N. Agueda, R. Vainio, B. Sanahuja, D. Lario An approach to the dynamics of extrasolar planets with satellites M. Andrade, J.A. Docobo Resonantly damped MHD waves in a system of two coronal loops I. Arregui, R. Oliver, J. Terradas, J.L. Ballester Coronal seismology using periods and damping rates of oscillating loops I. Arregui, J. Andries, T. Van Doorsselaere, M. Goossens, S. Poedts Circulation of Jupiter’s Polar Atmosphere from Cassini Images N. Barrado-Izagirre, A. Sánchez-Lavega, R. Hueso On the physical properties of Brown Dwarfs of the Chamaleon I Dark Cloud A. Bayo, D. Barrado y Navascués, M. Morales On the Direct Determination of Sensitivity, Resolution and Information Content of Helioseismic Data - Application to the Inversion of the Solar Core Rotation Rate A. Eff-Darwich, S. G. Korzennik, S.J. Jiménez-Reyes, R.A. Garcı́a Spectroscopy follow up of planet-bearing stars M.C. Gálvez, J. Ge, D. Montes Low-Mass Stars and Brown Dwarfs in NGC1333 C. Gómez Martı́n, E.A. Lada, A. Steinhauer, J.L. Levine Modelling Venus reflectivity in the visual range S. Pérez-Hoyos, A. Sánchez-Lavega Predicting the duration of Magnetic Cycles in Late-Type Stars R. Lorente, B. Montesinos Accurate properties of low mass stars: masses and radii of CM Draconis J.C. Morales, I. Ribas, C. Jordi, G. Torres, E.F. Guinan, D. Charbonneau Spitzer: An accurate variability study of late L dwarfs M. Morales-Calderón, J.R. Stauffer, D. Barrado y Navascués Search for eclipsing binary stars in open clusters with the ROTSEIIIb telescope I. Ribas, J.C. Morales, C. Allende Prieto, C. Jordi Homochirality: The Signature of Life J.M. Ribó, A. Moyano, A. Canillas, D. Hochberg, S. Veintemillas-Verdaguer Session IV. Observatories and instrumentation Contributed talks El Gran Telescopio Milimétrico/ The Large Millimeter Telescope E. Carrasco, the GTM/LMT collaboration BOOTES-IR: Simultaneous optical and near-IR observations with a 0.6m robotic telescope in Sierra Nevada A.J. Castro-Tirado, R. Cunniffe, A. de Ugarte Postigo, M. Jelı́nek, S. Vitek, P. Kubánek, J. Gorosabel, A. Riva, P. Conconi, F. Zerbi, A. Claret, S. Guziy, J.M. Castro Cerón, J.M. Trigo-Rodrı́guez, M.D. Sabau-Graziati Profiling of atmospheric turbulence: Activities at the Instituto de As- 362 A Appendix: Table of Contents of the CD-Rom trofı́sica de Canarias B. Garcı́a-Lorenzo, J.J. Fuensalida, M.A.C. -Rodrı́guez-Hernández, J.M. Delgado, C. Hoegemann Elmer performance after characterization and pre-shipping tests M.L. Garcı́a-Vargas, J.M. Martı́n-Fleitas, A. Cabrera-Lavers, J.M. Rodrı́guez-Espinosa, P.L. Hammersley, R. Kohley GRI: The Gamma-Ray Imager Mission M. Hernanz Interactive System for the Display and Analysis of Data Cubes Observed with Herschel D. Hurtado, E. Sánchez, J. Martı́n-Pintado The Gaia mission simulator X. Luri, E. Masana, P. Altamirano, Y. Isasi, C. Fabricius LIRIS: a new infrared spcetrograph on WHT A. Manchado, J. A. Acosta-Pulido, R. Barrera Radio Astronomy Data Model for Single-Dish Multiple-Feed Telescopes, and Robledo Archive Architecture J.D. Santander-Vela, E. Garcı́a, J.F. Gómez, L. Verdes-Montenegro, S. Leon, R. Gutiérrez, C. Rodrigo, O. Morata, E. Solano, O. Suárez The Alpha Magnetic Spectrometer on the International Space Station I. Sevilla, on behalf of the AMS collaboration Posters Near Infrared MOS using LIRIS J.A. Acosta-Pulido, R. Barrena, A. Manchado, C. Carretero, C. Domı́nguez-Tagle, A. Herrero, N. Rodrı́guez-Eugenio A. Vazdekis Simulations of gamma-ray detectors for astrophysics with Geant4 J. Alvarez, M. Hernanz EURECA: Towards high resolution X-ray spectroscopy X. Barcons, J. Bussons, M.T. Ceballos, J.R. Rodón, L. Fàbrega, F. Briones, J.V. Anguita, R. González, A. Camón, C. Rillo, F. Bartolomé, J. Sesé 3C A Multi-Wavelength database for the future GTC cosmological surveys G. Barro, J. Gallego, V. Villar, J. Zamorano OVO: Orion Virtual Observatory (in σ Orionis) J.A. Caballero Elmer Spectroscopy: Characterization and perfomance results from the pre-shipping acceptance tests A. Cabrera-Lavers, M.L. Garcı́a-Vargas, J.M. Martı́n-Fleitas, R. Kohley, E. SánchezBlanco, P.L. Hammersley, M. Maldonado, R. Vilela, J.L. Barreto, J.M. Rodrı́guezEspinosa Elmer Imaging: Characterization and perfomance results from the preshipping acceptance tests A. Cabrera-Lavers, M.L. Garcı́a-Vargas, J.M. Martı́n-Fleitas, R. Kohley, E. SánchezBlanco, P.L. Hammersley, M. Maldonado, R. Vilela, J.L. Barreto, J.M. Rodrı́guezEspinosa The GCS Data Processing Kit C. Garcı́a-Dabó, A. Cabrera-Lavers, P. Gómez-Cambronero A Appendix: Table of Contents of the CD-Rom 363 Science with OSIRIS H. Castañeda, J. Cepa, E.J. Alfaro, J.J. González, J.I. González Serrano, M. Sánchez-Portal, the OSIRIS team Scheduler Application to Transmit Real-Time Astronomical Reports S. Castillo-Carrión, A.J. Castro-Tirado, M. Jelı́nek, P. Kubánek Gamma-ray burst studies with BOOTES and BOOTES-IR in the it Swift Era A.J. Castro-Tirado, M. Jelı́nek, P. Kubánek, S. Vitek, A. de Ugarte Postigo, J. Gorosabel, R. Cunniffe, R. Hudec A sensitive all-sky camera for multipurpose recording of the nightly sky down to 10th magnitude A.J. Castro-Tirado, S. Vitek, P. Kubánek, M. Jelı́nek, J.M. Trigo Rodrı́guez, A. de Ugarte Postigo, T.J. Mateo Sanguino ICAT: a General Purpose Image Reduction and Analysis Tool for Robotic Observatories J. Colomé, I. Ribas Madrid Deep Space Communication Complex as Radio Astronomy Observatory C. Garcı́a-Miró, O. Morata, E. Moll, O. Suárez Development of VO services in Universitat de Barcelona F. Julbe, X. Luri, J. Torra The XMM-Newton Image Gallery after two years of life N. Loiseau, M. Ehle, R. Willatt, R. Bailey Cool stars in the solar neighborhood. Preparatory activities for the Darwin mission J. Maldonado, R.M. Martı́nez-Arnáiz, D. Montes, C. Eiroa, B. Montesinos, I. Ribas, E. Solano Popularizing activities of Astronomy at the Universitat de Barcelona E. Masana, S.J. Ribas, C. Jordi, J. Torra, High resolution spectra of cool stars in the Virtual Observatory. Criteria for spectral classification D. Montes, R.M. Martı́nez-Arnáiz, J. Maldonado, J. Roa-Llamazares, J. LópezSantiago, I. Crespo-Chacón, M.C. Gálvez, E. Solano, High resolution spectroscopic analysis of cool stars possible members of the AB Doradus moving group D. Montes, J. López-Santiago, I. Crespo-Chacón, R.M. Martı́nez-Arnáiz, J. Maldonado, PARTNeR: Radio astronomy for students O. Morata, O. Alles, O. Suárez Calar Alto: The Max Planck - CSIC Stage S. Pedraz The Gaia Telemetry CODEC: Results on data coding and compression J. Portell, E. Garcı́a-Berro, X. Luri, R. Castro, M. Prudhomme, Discovery of new candidate post-AGB stars and Planetary Nebulae with the Virtual Observatory M. Sierra, A. Bayo, P. Garcı́a-Lario, the AVO team The Gaia Consortium: Present and Future Spanish Contribution J. Torra, C. Jordi, X. Luri, F. Figueras 364 A Appendix: Table of Contents of the CD-Rom How to observe with GTC M.R. Villamariz Cid Index Abia, C., 239 Abreu, D., 81 Adami, C., 41 Aguerri, J.A.L., 131 Aguiar, M., 71 Alfaro, E.J., 71 Alonso-Herrero, A., 143 Anglada, G., 263 Arnaboldi, M., 41 Arnouts, S., 41 Asensio Ramos, A., 105 Bardelli, S., 41 Barreiro, R.B., 177 Barrera, S., 81 Barro, G., 209 Becerril, S., 81 Bellot Rubio, L.R., 271 Benı́tez, N., 185 Bland-Hawthorn, J., 71 Boehnhardt, H., 287 Bolzonella, M., 41 Bondi, M., 41 Bongiorno, A., 41 Bottini, D., 41 Busarello, G., 41 Cabrera-Lavers, A., 231 Cairós, L.M., 81 Cappi, A., 41 Castañeda, H.O., 71 Castander, F.J., 193 Cepa, J., 71 Charlot, S., 41 Ciliegi, P., 41 Cobos, F., 71 Comerón, F., 3 Contini, T., 41 Correa, S., 71 Cristallo, S., 239 Curiel, S., 263 Dı́az, J.J., 81 de Laverny, P., 239 Dierickx, P., 15 Domı́nguez, I., 239 Donley, J.L., 143 Eiroa, C., 279 Espejo, C., 71 Farah, A., 71 Fernández-Soto, A., 201 Foucaud, S., 41 Fragoso-López, A.B., 71, 81 Franzetti, P., 41 Fridlund, M., 279 Gómez, J.F., 263 Gago, F., 81 Gallart, C., 247 Gallego, J., 209 Garfias, F., 71 Garilli, B., 41 Garzón, F., 81, 231 Gavignaud, I., 41 Gigante, J.V., 71 González, C., 81 366 Index González, J.J., 71 González-Escalera, V., 71 González-Fernández, C., 231 González-Serrano, J.I., 71 Gorosabel, J., 29 Grange, R., 81 Gutiérrez, P.J., 287 Guzmán, R., 51 Guzzo, L., 41 Hammersley, P.L., 231 Hernández, B., 71 Herrera, A., 71 Hueso, R., 303 Ilbert, O., 41 Iovino, A., 41 Kadler, M., 165 Kaltennegger, L., 279 Kaufmann, S., 165 Koo, D.C., 209 Kovalev, Y.Y., 165 López, P., 81 López-Corredoira, M., 231 Lamareille, F., 41 Lara, L.M., 287 Lario, D., 295 Le Brun, V., 41 Le Fèvre, O., 41 Maccagni, D., 41 Manrique, A., 157 Marano, B., 41 Marinoni, C., 41 Martı́nez, M., 321 Martı́nez-González, E., 177 Martı́-Ribas, J., 219 Mathez, G., 41 Mazure, A., 41 McCracken, H.J., 41 Mellier, Y., 41 Meneux, B., 41 Merighi, R., 41 Merluzzi, P., 41 Militello, C., 71 Noeske, K., 209 Oscoz, A., 331 Pérez, J., 81 Pérez, R., 71 Pérez-González, G., 143 Pérez-González, P.G., 209 Pérez-Hoyos, S., 303 Paltani, S., 41 Pascual, S., 209 Patel, N.A., 263 Patrón, J., 81 Pelló, R., 41 Peraza, L., 71 Picat, J.P., 41 Pollo, A., 41 Pozzetti, L., 41 Radovich, M., 41 Rasilla, J.L., 71, 81 Redondo, P., 81 Restrepo, R., 81 Rieke, G.H., 143 Rigby, J.R., 143 Ripepi, V., 41 Rizzo, D., 41 Rodrı́guez-Espinosa, J.M., 63 Ros, E., 165 Rubiño-Martı́n, J.A., 177 Sánchez, B., 71 Sánchez, V., 81 Sánchez-Blázquez, P., 117 Sánchez-Lavega, A., 303 Sánchez-Portal, M, 71 Saavedra, P., 81 Scaramella, R., 41 Scodeggio, M., 41 Solano, E., 345 Stankov, A., 279 Straniero, O., 239 Tejada, C., 71 Telesco, C.M., 91 Tenegi, F., 81 Torra, J., 255 Torrelles, J.M., 263 Tresse, L., 41 Trujillo Bueno, J., 311 Tueller, J., 165 Index Vallbé, M., 81 Vettolani, G., 41 Vicente, B., 231 Villar, V., 209 Weaver, K.A., 165 Zamorani, G., 41 Zamorano, J., 209 Zanichelli, A., 41 Zucca, E., 41 367

Highlights of Spanish Astrophysics IV

Related documents

Products

Support

Highlights of Spanish Astrophysics IV

Related documents

Add this document to collection(s)

Add this document to saved

Suggest us how to improve StudyLib